The FCC says all of Iowa has access to broadband internet. Speed tests tell a different story.

According to FCC, Iowa is the only Midwestern state with virtually complete access to high-speed internet. Every county is covered by download speeds of 25 megabits per second (Mpbs), which the agency defines as “baseline” broadband. But another set of data tells a different story: Internet users in Iowa experience that speed only 22 percent of the time. That’s according to nearly half a million speed tests run on a diagnostic tool operated by the Open Technology Institute, a research arm of the New America Foundation, a non-partisan think tank. Data from these tests, which were run last year, mostly as Google searches, are publicly available on the institute’s website and were updated at the request of The New Food Economy.

Connections are worse outside the cities. Take, for example, a cluster of predominantly rural counties located in southern Iowa, between Des Moines and the Missouri border. According to the FCC map, 100 percent of residents in Appanoose, Davis, Lucas, Marion, Monroe, Wapello, and Wayne counties have access to a download speed of 25 Mbps, the speed at which three people can simultaneously stream HD video. But tests run in these counties show that can only happen 17 percent of the time.

On the outskirts of Albia, the Monroe county seat, wedding photographer Carol Selvy can’t show clients her photos. Selvy says she needs a week to upload a suite of files, which progresses in fits and spurts. According to FCC, Selvy has access to download speeds of 12 Mbps. (The agency’s “minimum” broadband definition is 10 Mbps.) But that isn’t accurate. Selvy played The New Food Economy a voicemail left by a representative from Windstream, a phone company with a large presence in Iowa, and her internet service provider.

“The best you’re gonna get is 4 megs,” the representative told her. That was the broadband standard in 2011.

Selvy is not alone. Corn farmers in Appanoose and Monroe counties are among the state’s least productive. Some say that’s due, in part, to slow connections. “We’ve had some of our members that want to take advantage of precision planting and spraying, but they just don’t have access to high-speed internet to be able to do it,” says Bryon Stilley, CEO of Chariton Valley Electric Cooperative, a member-owned utility that serves the counties. Because DSL connections in this part of Iowa are slow and satellite service is unreliable, many of the cooperative’s members rely on cell phone hotspots to get online.

Why do speed tests conflict so dramatically with what’s on FCC’s map? Because the broadband map, which the Commission calls a “key source” of information for consumers and policymakers, doesn’t include on-the-ground measurements in the first place. The map is based on data taken from Form 477, a filing that internet providers submit to FCC twice annually. The data are the agency’s main source of information on broadband availability, and the backbone of its funding decisions.

Form 477 data have surprising limitations. Providers are not required to include information in the filing about actual on-the-ground internet speeds, which are confidential and considered a trade secret. Instead, when providers submit data, they include lists of census blocks where they “can or do” offer service to at least one location, along with the maximum speeds they advertise there, whether that’s what residents have or not. Nationwide, around 28 people live in a census block, on average. In Iowa, a rural state, the density is closer to 15.

For these reasons, it’s hard to know how many Americans covered in the federal broadband map actually have the internet their providers say they do. When asked why FCC accepts self-reported advertised speeds as a metric, Mark Wigfield, a spokesman, said carriers are “required by statute to provide accurate data,” and the agency finds a strong correlation between advertised and actual speeds, though not for DSL or satellite. “The cost and burden of collecting actual speeds would be too great,” he wrote in an email to The New Food Economy.

Policymakers have openly acknowledged that the broadband map, the first major refresh of publicly accessible internet data in several years, is incorrect. When the map was unveiled in February, Commissioner Jessica Rosenworcel, who lives in Washington, D.C., admitted it had “errors,” and asked the public to email FCC directly with corrections. “I looked up my house and can tell you with good authority it lists service that is not available at my location,” she wrote. “You can go ahead and plug in your address and you might find the same thing.”

Part of the reason for this discrepancy is that FCC doesn’t collect granular data about deployment. “For a long time, the way that the FCC collected data about broadband was, as we found, if there is one subscriber in a census block, we presumed that it was available throughout the block,” Rosenworcel told a House subcommittee in 2017. “I think we all know that that is not a fair assumption anymore, and we’re leaving too many households behind.”

And faulty data can have real-life consequences. When census blocks are reported to have access to 10 Mbps downloads, which was the broadband standard in 2014, other internet service providers are disqualified from receiving FCC funds to expand service there. In other words, the government decides the area is connected enough not to require additional funds for expansion.

But self-reported estimates don’t match the speeds independently measured by the Open Technology Institute in Iowa, and the circumstances there are likely to be illustrative of a larger problem. Interviews with broadband advocates, public comments submitted to FCC, and statements by policymakers all indicate that in other states, rural residents may be in similar situations: reliant on subpar DSL and satellite connections, and ineligible for federal support.

Across the country, politicians with rural constituents are moving to address the problem. Last year, Republican Senator Roger Wicker of Mississippi, Democratic Senator Joe Manchin of West Virginia, and Democratic Congressman Dave Loebsack of Iowa all introduced bills to improve the accuracy of mobile broadband data—also from Form 477—in advance of an FCC funding opportunity for phone companies. Loebsack’s bill was signed into law in March as part of the omnibus spending bill.

“I like to call it, ‘garbage in, garbage out,’” Loebsack told telecommunications executives at a House hearing in January. “If we don’t have accurate data, whether that’s in agriculture, or that being a subset of something larger, then we’re not going to be able to make good public policy decisions, or even investment decisions on the part of the private sector.”

Manchin has been more direct. One month after introducing his bill, he admonished FCC Chairman Ajit Pai for the agency’s poor record-keeping. “Inaccurate data has failed rural and remote communities across this country,” he said at a Senate hearing. “Inaccurate data has caused us to be left behind.”

More recently, Loebsack, who represents southern Iowa, introduced another bill, co-authored by Republican Congressman Bob Latta of Ohio, that directs FCC to identify broadband coverage gaps on farms, and target funding to help farmers use data-heavy precision agriculture tools. (The bill passed a House subcommittee last Thursday.) But for the residents in his district, and particularly those who had hoped to benefit from an upcoming, billion-dollar broadband subsidy auction, the bill comes too late.

Bryon Stilley from the Chariton Valley Electric Cooperative is one of those residents. In 2016, he began planning a fiber broadband network for the 6,100 members of the co-op. Then, as now, rural members were buying internet from one of four companies: Windstream, the local phone company, which offered DSL; Rise Broadband, a “fixed wireless” provider that transmitted a wireline connection from towers; and two satellite companies named HughesNet and ViaSat. Disappointed by the options, many instead relied on service from their cell phones.

Generally speaking, there’s not a business case for bringing high-speed internet to rural America. It’s labor-intensive, expensive and inefficient. “We’re talking four-point-two, four-point-three members per mile of line,” Stilley says of his area. “At the end of the day, it’s hard to justify, and hard to make something like that work.”

To build a $29.3 million, 1,400-mile fiber optic cable network, the cooperative would need help. Stilley decided to pursue subsidies from FCC and its primary vehicle for funding rural broadband, known as the Connect America Fund. In the past, all of these funds went to large telephone companies, such as AT&T and Verizon, and in Iowa, to CenturyLink, Frontier and Windstream. Next month, the Commission will begin offering funds to small providers, including electric cooperatives.

That move comes too late for Stilley, whose plans to take advantage of Connect America Fund monies were dashed in December, when FCC updated a map of eligible areas and showed a major broadband expansion in his area. According to Form 477 data, Windstream and Rise Broadband had blanketed Chariton Valley’s service area in minimum broadband, leaving only nine census blocks uncovered and thus eligible for funding. Across the state, Rise had doubled its coverage. Possible funds for the cooperative shrank from $8 million to $56,000, Stilley says.

But speed tests don’t show an increase in high-speed internet. Last year, internet users in the seven counties in Chariton’s service area were just as likely to reach the minimum broadband speed of 10 Mbps as they were three years ago: around 44 percent of the time, according to the Open Technology Institute. Windstream users reached minimum broadband in only 34 percent of tests. Surprisingly, of thousands of tests conducted in those counties in 2016 and 2017, not a single user was on Rise Broadband, calling into question the efficacy of its advertising.

In an email to The New Food Economy, a Windstream spokesman said the company had expanded broadband coverage through VDSL bonding, a process of releasing additional, usually pre-existing telephone lines for internet use. Meanwhile, a spokesman for Rise Broadband said the company “has not had any significant Iowa changes,” and “coverage has remained about the same,” also in an email. Both companies use a mix of advertising, including direct mail and digital ads, to promote their services in southern Iowa.

Dependence on DSL is common in rural America, according to Brian Whitacre, an agricultural economics professor at Oklahoma State University and a professional bodybuilder. In cities, consumers often have their choice of wireline providers, which can include a local phone company offering DSL, but also cable and fiber companies. Not so in rural areas, where that phone company may be the only provider, and where internet connections degrade with distance.

Other rural technologies are unreliable. Fixed wireless, which is touted as “bridging the digital divide” in areas without wireline connections, requires a line-of-sight connection that can be interrupted by rain or trees. Satellite connections are also compromised by weather. One FCC study finds that satellite users rarely achieve advertised speeds. When FCC says 24 million Americans lack access to broadband internet, that includes internet users on these two kinds of connections.

In the case of Chariton Valley, it’s unclear whether broadband coverage actually increased, or was just reported to be more widely advertised.

Last year, FCC opened a proceeding to consider changing the Form 477 and seek granular deployment data, such as the locations of homes and businesses served, rather than census blocks reached. In seeking public comment, the agency wrote that filings do not have “meaningful information” about how data were collected or coverage areas determined. Nor had the agency investigated “whether actual consumer experience has diverged substantially from the Form 477 fillings.”

But FCC may never know the answer, because publicly subsidized providers don’t divulge their data collection process or explain how services are advertised in a specific market in their filings. They often claim that disclosing better, more specific deployment information would compromise customer privacy, and their ability to compete with other providers. (The Open Technology Institute supports collecting and disclosing more deployment data.)

Jonathan Chambers, a former FCC attorney and broadband consultant for rural electric cooperatives, including Chariton, believes that without an opportunity for his clients to challenge FCC’s data, the subsidy process will remain fundamentally flawed.

“If you’re gonna base the service—these public funding decisions—on data, you either have to account for the errors in the data, or you have to give people a chance to challenge the data,” Chambers says. “Because the funding decision means that these areas of the country now won’t get funded. And if they don’t get service, then they get the worst of both worlds. They don’t have service and they won’t have funding.”

Bryon Stilley says he is back to the drawing board. Now ineligible for Connect America Funds, Chariton Valley has applied for a grant from the Rural Utilities Services to build a smaller network in Appanoose, one of the seven counties in the cooperative’s coverage area. Absent the competition of another provider, it could be years before members in the other counties experience broadband internet.

Bidding for $1.98 billion in subsidies available through FCC’s Connect America Fund II auction, which has 277 applicants, including 25 rural electric cooperatives, begins July 24.