Case Study: Digital Redlining — Broadband Discrimination in American Cities

"The internet has become essential infrastructure. Denying it to communities of color isn't a market failure — it's a civil rights issue." — former FCC Commissioner Mignon Clyburn, 2020

Overview

In October 2022, The Markup, an investigative journalism nonprofit, published the results of a groundbreaking analysis: they had examined broadband internet offers available to 800,000 addresses in 38 major US cities. The findings confirmed what community advocates had long alleged — that internet service providers were systematically offering slower speeds at higher prices in neighborhoods with larger proportions of Black, Hispanic, and low-income residents. The practice had a name: digital redlining.

This case study examines the evidence for digital redlining, its causes, its consequences for data-driven equity, and the ongoing policy debate about how to address it. It connects directly to Chapter 32's central argument: that digital infrastructure inequality is not a historical problem being solved by market forces but an active, ongoing form of structural discrimination with compounding effects on every data system we have studied in this book.

Skills Applied: - Analyzing empirical evidence on infrastructure inequality - Connecting infrastructure access to algorithmic and data outcomes - Evaluating policy responses to structural discrimination - Applying the data justice framework to telecommunications policy

The Situation

What the Markup Found

The Markup's investigation, published as a series titled "The Divide," analyzed internet service offers from major ISPs — AT&T, Verizon, EarthLink, CenturyLink (now Lumen Technologies) — across 38 cities. For each address, researchers recorded the maximum available download speed and the monthly price. They then correlated these offers with the racial and income demographics of each neighborhood.

The findings were stark:

AT&T offered the least equitable service. In cities where AT&T was the primary provider, neighborhoods with the highest percentages of residents of color were offered internet speeds that were, on average, significantly slower than those offered to predominantly white neighborhoods — even when the neighborhoods were in the same city and had similar housing densities.

In Cleveland, AT&T offered speeds of 500 Mbps or higher to 76% of addresses in predominantly white census tracts but to only 13% of addresses in predominantly Black census tracts. The price difference compounded the speed gap: residents in lower-speed areas often paid comparable monthly rates for dramatically inferior service.

Verizon showed similar patterns. In New York City, the Markup found that Verizon's fiber-optic FiOS service — offering speeds of up to 940 Mbps — was significantly more available in wealthier, whiter neighborhoods in Manhattan and Staten Island than in predominantly Black and Hispanic neighborhoods in the Bronx and Brooklyn.

EarthLink and CenturyLink exhibited patterns consistent with the larger ISPs: slower speeds at comparable or higher prices in neighborhoods with higher proportions of minority and low-income residents.

Crucially, the Markup's analysis controlled for factors that ISPs typically cite to explain service disparities — housing density, building type, distance from network infrastructure. Even after these controls, the racial and income disparities persisted. The pattern was not explained by technical or geographic factors alone.

Historical Context: From Physical Redlining to Digital Redlining

To understand digital redlining, it is necessary to understand the physical redlining that preceded it.

In the 1930s, the Home Owners' Loan Corporation (HOLC) created maps of 239 American cities, grading neighborhoods on a four-point scale from "A" (Best) to "D" (Hazardous). The grades were explicitly tied to racial composition: neighborhoods with significant Black populations were almost universally graded "D" and marked in red — hence "redlining." These grades determined who could access federally backed mortgages, effectively denying Black communities the capital necessary to build wealth.

The Fair Housing Act of 1968 formally prohibited discriminatory lending, but the consequences of redlining persisted. The neighborhoods that were redlined in the 1930s remain, in many cities, the most economically disadvantaged neighborhoods today — with lower property values, less commercial investment, and less infrastructure of every kind.

Digital redlining maps onto this historical pattern with disturbing precision. A 2022 study by the National Digital Inclusion Alliance found that neighborhoods graded "D" by the HOLC in the 1930s are significantly more likely to lack broadband options today than neighborhoods that were graded "A." The infrastructure discrimination of the digital age follows the contours of infrastructure discrimination from nearly a century ago.

Key Actors and Stakeholders

Internet Service Providers

AT&T, Verizon, Comcast, and other ISPs control the physical infrastructure — fiber-optic cables, copper lines, cellular towers — through which broadband service is delivered. Their investment decisions determine which neighborhoods receive high-speed service and at what price.

ISPs argue that infrastructure investment decisions are driven by economic factors, not discrimination: deploying fiber-optic cable is expensive, and the return on investment is higher in neighborhoods with more affluent residents who can afford premium service. This argument treats market logic as racially neutral — but when market logic systematically produces racially disparate outcomes by building on a foundation of historical discrimination, the neutrality is illusory.

Local and State Governments

Municipal governments grant rights-of-way, franchise agreements, and permits that ISPs need to deploy infrastructure. Some cities have attempted to condition these agreements on equitable deployment requirements — mandating that ISPs serve all neighborhoods within a franchise area at comparable service levels.

However, state-level preemption laws complicate these efforts. Approximately 17 states have enacted laws restricting or prohibiting municipal broadband initiatives — often passed at the behest of ISP lobbyists who argued that government-owned networks would constitute unfair competition with private providers.

Community Organizations

Community technology organizations have been documenting and fighting digital redlining for years. The Detroit Community Technology Project, the National Digital Inclusion Alliance, and similar organizations provide community broadband, digital literacy training, and policy advocacy. These organizations often operate with limited resources but bring essential ground-level knowledge about the lived experience of digital exclusion.

Federal Regulators

The FCC has authority over broadband deployment and has periodically investigated digital redlining. In late 2022, the FCC launched a formal investigation into whether ISPs were providing equitable service across communities — a direct response to the Markup's investigation and years of community advocacy. The Bipartisan Infrastructure Law (2021) allocated $65 billion for broadband expansion, creating new resources and new questions about equitable distribution.

Analysis Through Chapter Frameworks

The Data Divide in Action

Digital redlining is not merely an access problem — it creates the data divide described in Section 32.1.3. When neighborhoods lack broadband:

1. Residents generate less data. Fewer online transactions, fewer digital health interactions, fewer social media posts, fewer search queries. The digital traces that algorithms depend on for training data are systematically thinner in digitally redlined communities.

2. Algorithms perform worse. Credit scoring models, health prediction models, employment recommendation systems — all depend on data that is less available for residents of underserved neighborhoods. The result is less accurate, often more biased, algorithmic decisions.

3. Services deliver less value. Google Maps is less accurate. Transit recommendations are less useful. Telehealth is inaccessible. E-commerce delivery is slower or unavailable. Each degraded service reduces the value proposition of broadband itself.

4. Reduced value discourages adoption. When the services available over broadband work poorly for your community, the incentive to pay for broadband diminishes — even if you can afford it.

This cycle means that digital redlining does not simply deny broadband to specific neighborhoods. It systematically degrades the entire data ecosystem for those communities, with cascading effects on algorithmic fairness, economic opportunity, and social participation.

The Compounding Effect

Chapter 32 describes how digital redlining compounds every algorithmic harm examined in the book. The case study evidence makes this concrete:

Healthcare. A patient in a digitally redlined neighborhood cannot access telehealth services, cannot use patient portals to manage chronic conditions, and cannot benefit from remote monitoring devices that require broadband connectivity. They are simultaneously underrepresented in the training data that health AI models depend on and unable to access the AI-powered health tools that could benefit them.

Education. A student in a digitally redlined neighborhood cannot participate in online learning, cannot access digital educational resources, and cannot develop the digital skills that the second-level digital divide requires. The 2020 pandemic made this visible when students in Detroit sat in fast-food parking lots to access WiFi — but the problem predated the pandemic and persists after it.

Economic opportunity. Job applications increasingly require online submission. Government benefits require online enrollment. Banking services migrate online. In each case, residents of digitally redlined neighborhoods face barriers that their counterparts in well-served neighborhoods do not — barriers that map onto and reinforce historical patterns of economic exclusion.

Environmental Justice Connection

Data center siting decisions intersect with digital redlining in troubling ways. Data centers — which power the digital services that digitally redlined neighborhoods cannot access — are often located in the same kinds of communities: areas with cheap land, less political resistance, and proximity to power plants. Communities may bear the environmental costs of data infrastructure (noise, water consumption, energy demand) while being denied the digital services that infrastructure enables.

The Policy Response

Federal Action

The Bipartisan Infrastructure Law (2021) allocated $42.45 billion to the Broadband Equity, Access, and Deployment (BEAD) Program — the largest broadband investment in US history. The program prioritizes "unserved" locations (those with no broadband access) and "underserved" locations (those with download speeds below 100 Mbps), with specific provisions for equitable distribution.

Additionally, the FCC's Affordable Connectivity Program (ACP) provided $30/month subsidies to low-income households for broadband service, reaching over 23 million households before its funding expired in early 2024. The expiration of the ACP demonstrated a recurring pattern: even when public programs address digital inequity, their sustainability depends on continued political will and funding.

Municipal Broadband

Some cities have built their own broadband networks to bypass ISP discrimination. Chattanooga, Tennessee, launched a municipal fiber-optic network in 2010 that offers speeds of up to 10 Gbps at competitive prices to all residents. The network has been credited with economic development, educational improvement, and elimination of digital redlining within the city.

But state preemption laws block this approach in many states. Tennessee itself enacted a law prohibiting Chattanooga from expanding its network beyond its existing electric utility service area — a law passed after intensive ISP lobbying.

Legal and Regulatory Action

In November 2022, the FCC adopted new rules requiring ISPs to display "broadband nutrition labels" — standardized disclosures of speed, price, and data caps — at the point of sale. These labels increase transparency but do not directly address the underlying infrastructure disparities.

Civil rights organizations, including the NAACP and the National Urban League, have filed formal complaints with the FCC alleging that ISP deployment patterns constitute illegal discrimination under federal civil rights law. The legal theory — that broadband infrastructure decisions that produce racially disparate outcomes violate Title VI of the Civil Rights Act of 1964 when the ISPs receive federal funding — remains untested in court as of early 2026.

Discussion Questions

1. Market logic and structural discrimination. ISPs argue that their investment decisions are driven by economics, not race. Evaluate this argument: Can a decision be economically rational and racially discriminatory at the same time? How does the concept of structural discrimination (as opposed to intentional discrimination) apply to broadband deployment patterns?

2. Broadband as infrastructure. The chapter and this case study argue that broadband should be treated as essential infrastructure — like water and electricity. What are the implications of this classification? Should broadband be subject to universal service obligations that require providers to serve all communities at comparable quality and price, regardless of profitability? What are the counterarguments?

3. The compounding effect. Choose one algorithmic system examined earlier in this book (predictive policing, credit scoring, health AI, content recommendation) and trace how digital redlining compounds its harmful effects on underserved communities. Be specific about the mechanisms — what data is missing, what decisions are degraded, and what alternatives are foreclosed.

4. State preemption. Approximately 17 states restrict or prohibit municipal broadband initiatives. Evaluate the ISP argument that municipal broadband constitutes unfair government competition with private enterprise. Consider: Is competition "unfair" when the private sector has failed to serve significant portions of the population? Whose interests do preemption laws serve?

Your Turn: Mini-Project

Option A: Local Investigation. Using the FCC Broadband Map, BroadbandNow, or comparable data, investigate broadband availability in two contrasting neighborhoods in your area. Document speed, price, and provider availability. Compare with demographic data (census tract-level data from data.census.gov). Write a one-page analysis of whether the patterns you find are consistent with digital redlining.

Option B: Policy Proposal. Draft a municipal policy that would address digital redlining in a specific city. Your proposal should include: (1) a universal service requirement for ISPs operating in the city, (2) an affordability program for low-income residents, (3) a digital literacy component, and (4) an accountability mechanism to monitor equitable deployment. Write a two-page policy brief.

Option C: Historical Mapping. Using HOLC redlining maps (available at mappinginequality.org) and modern broadband maps, compare the historical redlining grades of neighborhoods in a specific city with their current broadband availability. Write a one-page analysis examining the extent to which 1930s redlining maps predict 2020s broadband access.

References

• Ali, Christopher. "The Rise of Broadband and the Fall of Community: How Policy Has Undermined Rural Internet Access." Journal of Rural Studies 82 (2021): 285-295.

• Angwin, Julia, et al. "The Divide: America's Broadband Problem." The Markup, October-December 2022.

• Federal Communications Commission. "2023 Broadband Deployment Report." FCC, March 2023.

• Humphreys, Lee, et al. "Digital Redlining: The Role of Race in Internet Access." Information, Communication & Society 24, no. 5 (2021): 767-784.

• National Digital Inclusion Alliance. "Digital Redlining." NDIA Research Brief, 2022.

• National Telecommunications and Information Administration. "Internet Use Survey." NTIA, 2023.

• Noble, Safiya Umoja. Algorithms of Oppression: How Search Engines Reinforce Racism. New York: NYU Press, 2018.

• Pew Research Center. "Internet/Broadband Fact Sheet." Pew, 2023.

• Rothstein, Richard. The Color of Law: A Forgotten History of How Our Government Segregated America. New York: Liveright, 2017.