Closing Our Digital Divides

“Americans pay too much for Internet service,” President Biden recently declared.^a Broadband Internet access, in fact, has been a recurrent theme in several of Biden’s speeches, from announcing an Executive Order in 2021 to his most recent State of the Union Address. It is easy to see why the President would pay attention to Internet access in the U.S. We use Zoom, Teams, and Meet for work and school; we depend on Facebook and Twitter for news from friends, family, or society; TikTok, Netflix, and Spotify provide hours of entertainment. Medical appointments, government services, and commercial activity are all offered online. And broadband Internet access is the foundational service that undergirds all this economic and democratic activity.

Is Biden correct that the price is too high? Yes. While broadband Internet service providers (ISPs) dispute that characterization, recent studies find consumers are too often beholden to a monopoly provider, and the burdens of higher prices for comparatively lower-quality service fall, disproportionately, on historically disadvantaged populations. Hence, closing the “digital divide”—the gap between those communities that have reliable, affordable Internet access and those that do not—will require we study local broadband markets and muster the political will to address monopoly pricing and digital discrimination.

Back to Top

Is Broadband Too Costly?

It helps to begin with a primer on the dispute over broadband prices. Several commentators have echoed the president’s view that broadband Internet access is too expensive in the U.S., some suggesting that local monopolies are one cause of these high costs. Monopoly prices can put Internet access out of reach for many families—and U.S. residents without reliable and affordable broadband access also lack access to the commercial, educational, and civic opportunities that flow over the network.

For their part, broadband ISPs complain they do not get enough credit for keeping the bits moving. They say critics overlook how geographic differences affect broadband infrastructure costs. In their view, comparisons to international markets fail to account for the differences in, say, population density between South Korea and South Carolina. (It is generally more expensive to build a broadband network in a widely dispersed population than in a densely populated community.)

Similarly, when studies point out that prices are higher in local markets that lack competition, broadband ISPs respond that these studies unfairly group together far-flung, difficult-to-reach locations (that sometimes have fewer service providers), comparing them with those that are cheap and easy to serve (which usually have more options). In fact, under President Trump, the Federal Communications Commission (FCC) believed that so long as a broadband ISP like AT&T faces competition in some local market, we should treat it as though it faces competition in every market. The FCC explained that, in its view, so long as a broadband ISP had to compete for consumers’ business somewhere by reducing prices or improving service, the effects of that competition might spread across the entire company and be felt everywhere. (A federal court referred to these as “the benefits of competition” in “areas with fewer than two providers.”^b We usually just call those the problems of a local monopoly.)

Back to Top

Let’s Do the Numbers

Although there is disagreement on why broadband Internet access is so expensive—or even whether it is correct to call it “expensive” at all—there is a recent wave of studies aimed at examining, and closing, the digital divide. This work takes seriously the broadband ISPs’ complaints that geographic differences in, say, population density or topography matter, and so looks to variations within concentrated geographies. It finds that at least some communities are paying disproportionately higher prices for relatively lower-quality service.

One recent paper I co-authored surveys nearly one million addresses across 30 cities, comparing broadband ISPs’ offerings within each city.³ The results are striking, if not expected. Within those cities, the broadband ISPs’ various offerings are clustered around neighborhoods and communities. Some neighborhoods get great service: Fast download speeds at a good value. Consumers in other neighborhoods, however, get a raw deal: Slow (in some cases, practically unserviceable) speeds for the same—or higher—prices.

One key determinant of these outcomes is the presence of competition. Cox, for example, offers 30% faster download speeds (on a price-adjusted basis) in competitively served neighborhoods. In all, neighborhoods that can choose from among multiple broadband ISPs get better value, while those with “fewer than two providers” do not see any “benefits of competition.”

Importantly, not all competition is created equal. The study finds that cable companies such as Cox do not respond to “competition” from DSL at all. Put simply, where a broadband ISP’s only competition is DSL, it is as if there is no competition at all. That broadband ISP acts as a monopolist, serving slower speeds at higher costs. This finding is critical. As the U.S. federal government gears up to invest historic sums—more than $40 billion—in broadband infrastructure, some policymakers have expressed concerns about “overbuilding”—that is, funding new Internet access networks in regions where existing providers already offer service. Some providers, however, offer service that falls far short of modern bandwidth demands, and so competitors ignore it altogether. Policymakers should too. These findings confirm what has long been obvious to families living with limited options for broadband Internet access: We cannot rely on legacy or unreliable technologies to secure access to this defining modern utility.

Which families are caught on the far side of the digital divide? Unsurprisingly, recent research finds broadband ISPs offer some of their worst deals—slower speeds, higher costs—in regions home to many low-income families and communities of color. A team of investigative journalists at The Markup examined broadband ISPs’ offerings at more than 800,000 addresses in 38 cities.⁴ Their analysis, which controlled for population density, found that AT&T, for example, “disproportionately offered slower speeds to lower-income areas” in three-quarters of the cities studied.

Strikingly, The Markup found close correlations between broadband ISPs’ offerings and maps indicating the U.S. federal government’s sordid legacy of racism. In the 1930s, the Home Owners’ Loan Corporation (HOLC)—a government-sponsored corporation established to facilitate home ownership—created maps to assess the risks of funding mortgages in certain neighborhoods. HOLC graded neighborhoods with majority Black or African American residents as “high-risk” or “hazardous,” highlighting them in red on the agency’s maps. Such “redlining” has had long-felt effects in such matters as home ownership, household income, health outcomes, and, it now seems, broadband Internet access. According to the study, “redlined addresses were offered the worst deals almost eight times as often.”

AT&T called the study “fundamentally flawed” because it “ignored” the company’s “participation in the federal Affordable Connectivity Program” (ACP), which subsidizes discounts for low-income families.⁴ AT&T’s participation in ACP is beside the point. In these neighborhoods, AT&T seems to generally offer lower-quality service, and nothing about ACP fixes that. And while ACP lowers costs for eligible low-income families, it does not affect AT&T’s revenue. Whenever AT&T offers an ACP-related discount, the federal government uses public funds to reimburse AT&T. AT&T, then, is ACP-neutral: No matter whether a consumer benefits from an ACP discount, AT&T pockets the full charge. We should expect AT&T to offer the same higher-quality service in these communities as it does in higher-income neighborhoods.

Stark disparities thus characterize the digital divide. That divide exists not only in terms of bare Internet access; it also exists in terms of affordability and service quality. And the digital divide segregates us into digital-haves and digital-have-nots in ways that echo the legacy of our nation’s history of racism and class-based discrimination.

Back to Top

Bridging Divides

As important as it is to study and acknowledge these features of broadband Internet access in the U.S., it is far from enough. Policymakers in federal, state, and local governments must devise innovative policy and technical solutions to address the disparities in the availability and affordability of broadband Internet access.

One creative solution, out of New York, is to mandate a low-cost broadband Internet access option for low-income consumers. Specifically, New York requires that broadband ISPs offer a $15 per month package to eligible low-income consumers. And ISPs canot get away with selling subpar service in this low-cost tier: New York’s rules set minimum service standards in terms of download speeds. Rather than comply, broadband ISPs have sued New York, claiming that the state has no power to issue such Internet access rules. (States can in fact regulate local broadband services, but that is a topic for another time.²)

Stark disparities thus characterize the digital divide.

In this column, I address a different question: Whether such rate and service regulations are effective at addressing some of the problems described here. Again, the answer is yes.

While “rate regulation” is something of a taboo in some telecommunications policy circles, the fact is such rate controls are quite common. For example, the FCC grants public funds to private broadband ISPs to subsidize new network infrastructure in unserved, difficult-to-reach areas. But those grants come with strings attached: Broadband ISPs must use the grants to offer service “reasonably comparable” to what is available in other markets, and at rates that are “reasonably comparable” to those charged in other regions.

These regulations can improve consumer outcomes, sometimes dramatically.¹ Accounting, again, for geography by emphasizing comparisons within closely concentrated locales, I found consumers get better service, lower costs, or both as result of these regulations. Consider one example from Faribault, MN. At the time of my study, CenturyLink sold Internet access services there at a $49 flat rate. At some addresses, CenturyLink was an unregulated monopolist; at others, some only a five-minute drive away, CenturyLink was subject to the FCC’s regulatory conditions. Where CenturyLink’s service was regulated, consumers got download speeds of up to 40Mbps; where it could act as an unregulated monopolist, consumers got only 3Mbps.

These recent studies, and the policy approaches they examine, signal several important lessons for closing the digital divide. Regulators, alongside law and policy scholars, should draw upon the methods, tools, and data developed by the Internet measurement community and deployed in some of the studies noted above to analyze existing policy approaches and assess new possibilities. Likewise, Internet measurement scholars might turn to the law and policy academies to identify pressing questions in need of answers, or to policymakers to ask about their most urgent data needs. Closing the digital divide will require we bridge the gap between disciplines.

Such collaborations can help ensure our policymaking is informed by data and experience, not conjecture, offering a promising path to better and more affordable broadband Internet access. Indeed, interdisciplinary studies, including those described here, have shaped the FCC’s efforts to map broadband availability nationwide, influenced the National Telecommunications and Information Administration’s approach to funding new broadband infrastructure, and informed the White House’s call to address unaffordable broadband prices. Such studies suggest rate and service controls offer one promising path for addressing the persistent digital divide. Subsidies—both for broadband ISPs to expand access, and for consumers to improve affordability—offer another. The market alone cannot do it, but some political will and a bit of policy creativity can help to close the wide gap between U.S. residents with reliable and affordable broadband Internet access, and those without it.