Before the

FEDERAL COMMUNICATIONS COMMISSION

Washington, D.C. 20554

 

In the Matter of        Implementing the Infrastructure + Investment and Jobs Act:  Prevention and Elimination of Digital Discrimination

) ) ) ) ) )

GN Docket No. 22-69

+ 

COMMENTS OF CEO ACTION FOR RACIAL EQUITY

CEO Action for Racial Equity (CEOARE) is pleased to submit our response to your Notice of Inquiry. We aim to provide input and recommendations for consideration to the Federal Communications Commission (FCC) in addressing digital discrimination through the bipartisan Infrastructure Investment and Jobs Act (IIJA).  We commend you for prioritizing the goal to ensure that every person in the United States has equal access to high-quality, affordable broadband internet access service as a modality to directly address racial disparities and improve the lives of historically marginalized and disadvantaged groups.

CEOARE is a Fellowship composed of over 100 companies that mobilizes communities of business leaders with diverse expertise, across multiple industries and geographies, to advance public policy in four key areas — healthcare, education, economic empowerment, and public safety. Our mission is to identify, develop, and promote scalable and sustainable public policies and corporate engagement strategies that address systemic racism, social injustice, and improve societal well-being.

The CEOARE Fellowship has a policy portfolio focused on eight issues that disproportionately and systemically impact Black Americans, including Closing the Digital Divide. Our comments are grounded in a principles-based approach, which has relevance across several topics within this Notice of Inquiry and which are of critical importance to Black Americans and communities of color, including:

• Accessibility: In urban and rural communities, access to a reliable broadband network will help enable full participation in society and strengthen the U.S. economy.
• Affordability: Even when broadband is available, it must be affordable for moderate to low-income Americans.
• Adoption: Barriers to technology adoption should be understood and addressed to help enable all to benefit from digital connectivity.
• Data Mapping and Tracking: Accurate data mapping is necessary to understand the intersection between racial equity and the Digital Divide.

With these principles in mind, we offer comments on several questions contained in the Notice of Inquiry, as outlined below. Specifically, we aim to address definitions of key terms, including “comparability” and “a given area,” as well as the use of available data to prevent and eliminate digital discrimination.

Comments on Prevention and Elimination of Digital Discrimination

13. We next seek comment on the two notions of comparability in the definition of equal access.  How should we understand the phrase “an offered service that provides comparable speeds, capacities, latency, and other quality of service metrics”?   What “other quality of service metrics” should we consider?  Should they change over time?  To the extent that the quality of service metrics change and evolve over time, how would we continue to judge comparability?  When considering these various metrics, what does it mean for speeds, capacities, latency, or other metrics to be “comparable”?  Should we establish a prescriptive range of differences within which a service would be “comparable” (e.g., a percentage difference)?  If so, how would we determine the appropriate range?  If not, how do we give meaning to this section while avoiding a prescriptive approach?  In the Universal Service context, Congress has charged the Commission with ensuring that rural and urban areas have “reasonably comparable” access to telecommunications and information services.   To implement this language, the Commission collects annual survey data of the fixed voice and broadband services offered to consumers in urban areas and uses this data to develop reasonable comparability benchmarks for eligible telecommunications carriers subject to public interest obligations for fixed broadband.   Carriers receiving high-cost support must certify annually that they are meeting these benchmarks.   In addition, carriers receiving support under our Connect America Fund and Lifeline Programs are required to meet applicable speed and latency standards and are subject to detailed performance testing requirements.   Are there any insights or lessons gleaned from how the Commission has employed this phrase in the Universal Service context that may be relevant to our interpretation of paragraph 60506(a)(2)?

We acknowledge the difficulty in setting defined parameters and quality of service metrics that directly address comparability for purposes of identifying and preventing digital discrimination.  In making this decision, it is important to consider how the needs of society change over time. For example, the COVID-19 pandemic has highlighted more than ever how essential broadband is to American life, particularly education, jobs, health and medical care. What were once acceptable standards of broadband service quickly became outdated, as more families were forced to share household broadband connections for the purposes of online education, remote work and job-seeking, and telehealth services. The Pew Research Center found that 29% of broadband users have had to improve the speed, reliability, or quality of their high-speed internet connection at home since the pandemic has started.^[1] Unfortunately, not all Americans have the ability to upgrade their broadband service appropriately.

The Digital Divide in America is heavily interwoven with issues of race, education, and economic status. Black Americans remain less likely than non-Hispanic white Americans to own a traditional computer or have high-speed internet at home.^[2] Black Americans are already ten years behind white peers in digital literacy and could be disqualified or underprepared for 86% of the jobs in the U.S. job market by 2045.^[3] If comparability metrics are not regularly evaluated to target this racial divide, Black Americans could fall even further behind.

Accordingly, we recommend the periodic reconsideration of comparability and quality of service metrics, including speed and capacity, to meet the societal needs of education, public safety, economic empowerment and healthcare. ​

14. How should our concept of comparable quality of service account for various technical practicalities? For example, how should we take into account the nature of network upgrade cycles, which may occur over a period of time?   How should we account for network outages, or periods of network degradation due to disruptions in service or high utilization?  Should we understand comparable quality of service to vary during times of network degradation?  How should we interpret comparability across different services, including evaluating fixed broadband versus mobile broadband services?  Should performance metrics be the same for fixed and mobile broadband?

Network stability, including outages and degradation, should be considered when developing comparable quality of service metrics for the purposes of identifying and preventing digital discrimination against Black Americans and communities of color. Network quality can vary significantly, especially in zip codes with predominantly minority groups average broadband speed that is 21.5% slower than average download speeds in primarily white/Asian zip codes.^[4]

In September of 2021, the Detroit community of Hope Village faced a 45 day internet and phone outage, an outage that was not resolved until the city stepped in to intervene.^[5] Hope Village is a predominantly Black neighborhood in a predominantly Black city where 33% of the population live below the poverty line.^[6] Likewise, in Houston, broadband service “becomes patchy between Interstate-69 and the Westpark Tollway” where the median household income is less than $30,000 a year.^[7] Yet, nearby wealthier neighborhoods receive more consistent coverage.^[8]

Accordingly, a standardized approach to evaluating quality of service that considers not just access, but quality of access, is necessary. Without such an approach, urban and low-income areas that are considered “served,” yet may still suffer the effects of outdated and degrading networks, could fall through the cracks, resulting in digital discrimination in communities of color.

17. Geographic Area. How should we construe the phrase “in a given area” in the definition of equal access?   Does the different word choice signify that this refers to something other than “the service area of a provider of such service” used elsewhere in this subsection?   If so, how should we construe the “given area” in which the “equal opportunity to subscribe” is called for?  What unit of geography would provide appropriate granularity and be easy to match with other data?  Should we interpret the phrase in such a manner that it would track established geographical lines, such as city, county, and state boundaries or general demographic data such as U.S. Census statistical areas?  Or should we define the “given area” in some way tied to the provision of broadband, such as the service area of a provider?  For example, are there common industry practices that divide the provisioning of broadband into definable areas that could be relevant here?  If “in a given area” should be construed as the service area of a provider, are ILEC service areas and local-franchise-agreement-defined areas for cable providers the right areas to use?  How should we approach entities that are not ILECs or cable providers?  Should we consider outage reporting metrics in our analysis?   Should we apply the same or different standards to define a “given area” for fixed and mobile services?  The Commission has found repeatedly that because most consumers use their mobile wireless services at or close to where they live, work, and shop, they generally purchase mobile wireless services from service providers that offer and market such services locally.   Should we adopt the geographic market definition used in wireless transactions that defines and examines a local area, or would a larger geographic area be more appropriate?

It is critical that a “given area” be granular enough to protect historically marginalized groups who primarily live in urban areas.  U.S. cities and urban counties have many residents who lack access to home broadband service. Specifically, 13.9 million metropolitan households live without an in-home or wireless broadband subscription.^[9] For comparison, this is more than triple the 4.5 million rural households without a broadband subscription. ^[10]  

When digging deeper from the city/county level to the neighborhood level, many low-income neighborhoods have a dramatic drop in internet service access as compared to wealthier neighborhoods just a few miles away.^[11]  For example, some public housing units were not historically considered when initial broadband infrastructure was constructed and continue to lack reliable service today.^[12] Zip codes with predominantly minority groups average broadband speed that is 21.5% slower than average download speeds in primarily white/Asian zip codes.^[13] Therefore, if a “given area” is limited to markets at even the city or county level, many urban residents may not be afforded the digital discrimination protections as outlined in the IIJA.  

Accordingly, we recommend that a geography for a “given area” should be defined as no larger than a zip code in order to provide sufficient safeguards to communities of color and low-income areas.

23. We seek comment on the listed characteristics of “income level, race, ethnicity, color, religion, or national origin[.]” We recognize that many of these terms have established meanings in other areas of law regarding discrimination. Do we need to further define these terms, or is their meaning self-evident, especially in light of existing precedent?  If we did further define these terms, should we defer to other precedents or sources of law to give them meaning and, if so, which?  Should our efforts to prevent digital discrimination focus on preventing discrimination against particular individuals or communities in the aggregate that meet one of the listed characteristics?  If we focus on communities in the aggregate, how do we determine that a given community meets the listed characteristics?  Should we look to demographic data or standards, such as those used by the Office of Management and Budget (OMB)—which we have followed in other contexts —or the Census Bureau?   How should we account for the fact that data sources may not use the specific characteristics listed in paragraph 60506(b)(1) in their data collections?  For example, the Census Bureau does not use the term “color” in its data collections; how can we ensure that we benefit from robust use of existing data while following the language of the statute?

We urge the Commission to adopt definitions of protected classes based on prior legislation, regulations, and precedent, including the Civil Rights Act of 1964. As terms such as race and ethnicity have been accepted in the national lexicon, it will be easier for individuals who allege digital discrimination to know if they are members of a historically protected class.

While preventing digital discrimination based on protected characteristics at the individual level is key to helping close the Digital Divide, we urge the Commission to also consider claims by communities in the aggregate. For example, many predominantly minority neighborhoods were ignored when broadband infrastructure was constructed, resulting in slower speeds and access challenges for these communities.^[14] By accounting for digital discrimination at the community level, established analytical practices can be conducted to determine: 1) whether a statistically significant number of community members are in a protected class; and 2) whether there is evidence of disparate impact, even if the practices and policies by entities in question are facially neutral, as established by prior precedent (e.g., employment and housing discrimination).

25. To what entities should our rules preventing digital discrimination apply? The Commission has previously recognized that consumers can access broadband through a range of technologies, such as digital subscriber line (DSL), cable modem, fiber, wireless, and satellite, and that broadband can be fixed or mobile. Can the providers of all of these types of broadband engage in digital discrimination?  Can entities other than broadband providers engage in digital discrimination and, if so, what types of entities?  For example, can owners of multiple-tenant environments digitally discriminate against those living and working in their buildings?

We recommend that “entity” should be defined broadly, as there are any number of actors that can influence broadband access, affordability and adoption, from governments down to landlords. Because digital discrimination can be both a systemic issue and an individual actor issue, it is imperative to consider the entire broadband logistics network in order to adequately dissuade discriminatory practices.

26. Identifying Instances of Digital Discrimination. We seek comment on how to identify when and where digital discrimination is occurring. In order to identify areas and individuals impacted by digital discrimination, we will benefit from the use of data.  We seek comment on data sources that would enable us to identify occurrences of digital discrimination based on the listed characteristics.  For example, would data regarding demographic characteristics and broadband availability and adoption information be of particular importance to this analysis?  If so, what sources should we rely on for data regarding broadband information and demographic characteristics?  Are there other categories of data that are fundamental to such analysis?

In order to properly address digital discrimination against Black Americans and communities of color, any data used should be disaggregated by race. Data that is not disaggregated or is disaggregated only by general buckets of protected characteristics (e.g., women and minorities combined) does not tell a complete story of where gaps may exist in areas of access, affordability, and adoption of broadband services. Additionally, thorough analysis of a disparate impact claim of digital discrimination cannot be completed without disaggregated data by race.

For example, based on disaggregated racial data from the U.S. Census Bureau,^[15] the proportion of Black households that have access to computers and internet connectivity are uniformly and disproportionately lower, compared to state averages and other racial/ethnic groups, with Black households being, on average, five percent less likely to have access than the state averages across the United States.^[16]

Accordingly, we urge the Commission to prioritize disaggregated data by race to sufficiently protect Black Americans and communities of color from digital discrimination.

27. Could we leverage existing and pending Commission information collections regarding broadband to support identification of areas and individuals potentially impacted by digital discrimination? For example, the Broadband Data Task Force is currently leading an effort to collect location-based data on fixed broadband service availability and develop more precise maps through the Broadband Data Collection. Will the Broadband Data Collection data—indicating the availability of broadband, subject to challenge by stakeholders, with breakouts by technology, and download and upload speed—be an appropriate tool, in combination with other data, to help evaluate whether access is equal?  What kinds of analyses would be appropriate in making that determination?  How could we account for network upgrade cycles, which may occur over an extended period and may not be reflected in data that represent a given point in time?  Are there other sources of broadband availability information that the Commission should look to when conducting analyses?  Will relying on data sources that are not specifically designed to identify discrimination assist the Commission in effectively identifying such areas?  Would data on broadband subscriptions or adoption be relevant in evaluating the presence of digital discrimination?  The FCC collects tract-level, fixed-broadband subscription data and state-level, mobile-broadband subscription data on its Form 477; and the Census Bureau includes questions on broadband subscriptions in its American Community Survey.  How could such data assist in identifying instances where access is not equal?  Should the Commission, either instead of or in addition to relying on existing data sources, consider a new data collection?  If so, what information should be collected to identify and measure discrimination in quality of service, deployment, reliability, and other forms?   For example, should the Commission collect additional information related to the reliability of broadband networks in the Commission’s Network Outage Reporting System (NORS) and Disaster Outage Reporting System (DIRS) and if so, what specific information should be collected?  Should the Commission collect data on usage caps, throttling, and speeds that the service was reduced to?  If so, what is the appropriate frequency to gather this data and should it be at the subscriber level?  Would it be necessary to collect subscriber-level information and are there any legal or practical obstacles to doing so?   If not collected at the subscriber level, what would be an appropriate level of granularity for such a data collection?  Would any such collections present unique privacy concerns and, if so, how should we address them?

We affirm that the Broadband Data Task Force has a significant opportunity to close existing data gaps by collecting and compiling location-based data on fixed broadband service availability and developing more precise maps. Currently, Form 477 is self-reported by providers with little oversight, increasing the risk of inaccurate tracking and reporting.  For example, Form 477 data shows 15M Americans are disconnected from broadband compared to Microsoft-sourced data, which shows 120M Americans are disconnected.^[17] This substantial difference highlights the need for more accurate data collected at the federal level.

Similarly, the annual American Community Survey by the U.S. Census Bureau uses a relatively small sample population of 3.5M addresses per year,^[18] leaving open the possibility of data gaps, especially in low-income areas where housing transience is high. “Low-income households have experienced higher-than-usual levels of housing insecurity during the pandemic. Transience deters people from investing time in setting up internet service.”^[19] Housing transience also reduces the likelihood that a survey sent to a given address will be completed.

Accordingly, we encourage the Broadband Data Task Force to look to outside data to fill in gaps from federally collected information. This could include data from public-private relationships that specifically account for race so that digital discrimination can be appropriately targeted to increase impact in assisting communities of color.

28. Could we leverage data sources outside of the Commission for alternative or additional data? For example, the United States Census Bureau has existing demographic surveys, including the decennial census, some of which are conducted on an ongoing basis and measure many of the demographic characteristics specified in section 60506(b)(1). Are there particular surveys conducted by the Census Bureau that would be appropriate to rely on?  What survey design aspects are most important for the Commission to consider when determining whether a dataset is suitable for such analysis?  In particular, what is an appropriate level of geographic granularity for such information (e.g., census tract, block group) and how should the Commission balance the tradeoff between granularity and recency of the data?  In addition, would data from the Census Bureau (e.g., non-public, more disaggregated or raw survey data) be useful to the Commission in identifying instances of unequal access?  We seek comment on any other sources of demographic information and factors the Commission should consider in identifying data to rely on.  Would using any data sources present unique privacy concerns and, if so, how can we address them?  We also invite commenters to identify instances of digital discrimination, supported with data where possible, to help us better understand the scope and nature of digital discrimination.  We acknowledge the work that has been done to document these issues, and invite commenters to address how they should inform the Commission’s efforts.

We envision that public-private data could be used to supplement federal data. For example, public-private partnerships like Microsoft, Open Data Institute, and BroadbandNow have collected considerable broadband data and created tools for in-depth data analysis, including by race and zip code. A recent study by Pew Research shows that thirty-six states have public broadband maps showing service availability for their state. Many of these maps were forged by public-private partnerships.^[20]

Additionally, the Commission can continue to promote self-collection methods like the FCC speed test app (crowdsourcing) and the challenge process outlined in the Broadband Data Act of 2020,^[21] so that individual users can continue to contribute to close data gaps and disparities.

36. Apart from the Commission’s existing informal consumer complaint process, should we establish an alternative complaint process for violations of any rules we adopt to prevent digital discrimination? Many anti-discrimination laws and frameworks enable individuals to bring individualized complaints. Would such a scheme be practicable and desirable in the context of digital discrimination at issue here?  How would it work, what would be the requirements to make a successful claim, and what remedies would be available to individuals who make a successful claim?  Is there an existing alternative complaint process that the Commission could look to in developing a process for accepting complaints related to digital discrimination?  Should we establish a dedicated ombudsperson to use alternative dispute resolution to facilitate resolution of such complaints?

As the Commission considers what type of complaint process is most appropriate for complaints of digital discrimination, we recommend that the Commission consider examining the practices of other federal agencies that address potential discrimination. In addition, we recommend that the Commission take into account the perspectives of underserved communities in deciding the process to promote accountability on digital discrimination. Finally, we recommend leveraging leading practices around stakeholder engagement from digital inclusion leaders, such as the Black Brilliance Research Project and the National Digital Inclusion Alliance, for purposes of developing a complaint process that is informed by the very individuals it seeks to protect. Examples of such leading practices to obtain stakeholder perspectives on issues of digital equity are noted below:^[22], ^[23]

• The use of funding mechanisms to encourage participation in feedback sessions and follow-up research in acknowledging the voices and time of underserved community members.
• Encourage the submission of data-driven, anecdotal, and/or aggregate feedback on both individual service issues, as well as more systemic challenges. This could include community testing of browser-based and SMS-based speed tests, as well as verbal commentary through testimonies in broadband town halls. Opportunities for community members to provide feedback to civic leaders, elected officials, and government departments in order meet community-identified needs related to IIJA goals are important to understand and address systemic issues.
• Foster accountability and systemic change by employing such tactics as participatory action research, which could be key to planning, rollout, and follow-up evaluations in response to community feedback and points of view.

Conclusion

We appreciate that the FCC is looking for opportunities to prevent and eliminate digital discrimination through the Infrastructure Investment and Jobs Act. We hope our responses convey our enthusiastic interest in advancing racial equity priorities that seek to close the Digital Divide.

+

Respectfully submitted,

CEO Action for Racial Equity

---

Citations

[1] Colleen Mcclain, Andrew Perrin, Lee Rainie, Stella Sechopoulos, and Emily A. Vogels, “The Internet and the Pandemic”, Pew Research Center, Washington, D.C., September 1, 2021.

[2] Sara Atske and Andrew Perrin, “Home Broadband Adoption, Computer Ownership Vary by Race, Ethnicity in the U.S.”, Pew Research Center, Washington, D.C., July 16, 2021.

^[3] Apjit Walia, “America’s Racial Gap & Big Tech’s Closing Window”, Deutsche Bank Research, September 2, 2020.

[4] Infrastructure Bill Update, Tufts Fletcher School Digital Planet, February 2022.

[5] Malachi Barrett, “Detroit’s Plan for Equal Internet Access Starts with Hope Village”, BridgeDetroit, April 14, 2022.

[6] QuickFacts: Detroit City, Michigan, U.S. Census Bureau, July 1, 2021.

[7] Lara Fishbane and Adie Tomer, “Allan Holmes, Eleanor Bell Fox, Ben Wieder and Chris Zubak-Skees, “Rich People Have Access to High Speed Internet; Many Poor People Still Don’t”, The Center for Public Integrity, May 12, 2016.

[8] Id.

[9] Lara Fishbane and Adie Tomer, “Neighborhood Broadband Data Makes it Clear: We Need an Agenda to Fight Digital Poverty”, The Brookings Institution, February 6, 2020.

[10] Lara Fishbane and Adie Tomer, “Allan Holmes, Eleanor Bell Fox, Ben Wieder and Chris Zubak-Skees, “Rich People Have Access to High Speed Internet; Many Poor People Still Don’t”, The Center for Public Integrity, May 12, 2016.

[11] Id.

[12] Id.

[13] Infrastructure Bill Update, Tufts Fletcher School Digital Planet, February 2022.

[14] Lara Fishbane and Adie Tomer, “Allan Holmes, Eleanor Bell Fox, Ben Wieder and Chris Zubak-Skees, “Rich People Have Access to High Speed Internet; Many Poor People Still Don’t”, The Center for Public Integrity, May 12, 2016.

[15] 2018 American Community Survey, U.S. Census Bureau.

[16] Digital Injustice, The Fletcher School at Tufts University, June 22, 2020

[17] Microsoft Broadband Maps, October 2020.

[18] American Community Survey Information Guide, U.S. Census Bureau, October 2017.

[19] Chris Goodchild, Hannah Hill, Matt Kalmus, Jean Lee, and David Webb, “Boosting Broadband Adoption and Remote K–12 Education in Low-Income Households”, Boston Consulting Group, May 12, 2021.

[20] Lily Gong and Anna Read, “Which States Have Dedicated Broadband Offices, Task Forces, Agencies, or Funds?”, The Pew Charitable Trusts, November 30, 2021.

[21] 47 U.S.C. § 641 et seq.

[22] Black Brilliance Research Project, Comments on Implementation of the Infrastructure Investment and Jobs Act, Docket No. 220105-0002, February 3, 2022.

[23] National Digital Inclusion Alliance, Comments on Implementation of the Infrastructure Investment and Jobs Act, Docket No. 220105-0002, February 4, 2022.