Racial/Ethnic and Income Disparities in Neighborhood-Level Broadband Access in 766 US Cities, 2017- 2019

doi:10.21203/rs.3.rs-1662990/v1

PDF

Article

Racial/Ethnic and Income Disparities in Neighborhood-Level Broadband Access in 766 US Cities, 2017- 2019

https://doi.org/10.21203/rs.3.rs-1662990/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Broadband access is an essential social determinant of health, the importance of which has increased due to the ongoing COVID-19 outbreak. Using data from the City Health Dashboard and American Community Survey, we found that in 2017, 30% of urban households in 766 large US cities still had no access to high-speed broadband internet. After controlling for household income, broadband access in majority Black and Hispanic neighborhoods was 10-15% lower than in White or Asian majority neighborhoods. In 2019, lack of broadband access in urban households decreased to 28%, but substantial broadband disparities still persist in these cities.

Access to fast, reliable internet (“broadband”) is an important resource for obtaining essential services and information, including for health. Broadband is widely used to access health-related resources, including remote monitoring medical devices, delivery of healthy food from web-based food services, and medical information.¹ Conversely, lack of broadband hinders access to essential services.^2,3 The ongoing COVID-19 pandemic has exacerbated the consequences of lack of broadband access, raising awareness of its importance, especially for racial/ethnic minority populations and in rural communities.^1,4 Despite the growing importance of broadband in daily life, a substantial number of US residents still lack access.²

Researchers have characterized urban-rural differences in internet access.^2,5 Rural communities are more likely than urban communities to suffer from lack of any type of internet access. In 2018, 19% of rural households still had no form of internet, compared to 14% of urban households. Even fewer households had access to high-speed broadband internet services.⁶ Many studies have attributed these disparities to the lack of investment and deployment of broadband infrastructure in rural areas.^2,5 In the 2021 infrastructure bill, $65 billion dollars were allocated to build broadband infrastructure in rural areas, a step that should help to ameliorate infrastructure as a barrier to connection.⁷ Yet, while most U.S. cities now have the physical infrastructure to support broadband access for all residents, substantial disparities in broadband access persist within cities and across neighborhoods. Broadband gaps in urban environments are largely influenced by lack of affordability, disparities in digital literacy, as well as higher difficulties of broadband access among populations with lower educational attainment and language barriers.^8,9 These factors suggest that infrastructure improvements alone may not be sufficient to eliminate disparities in broadband access for urban households.¹⁰

To characterize within-city broadband access disparities, which have received less research attention than urban-rural broadband access disparities, we examined disparities in broadband access in 766 US cities by contrasting median population-level broadband internet access across cities and at the neighborhood level (proxied by Census tracts) using the data from American Community Survey (ACS), 2017-2019. To understand racial/ethnic and income disparities in broadband access, we stratified by neighborhood racial/ethnic composition and median household income. The goal of this analysis was to provide policymakers and researchers with a clear understanding of within-city disparities in broadband access, in order to identify potential solutions to increase access and utilization, and ultimately improve access to health-related resources in US cities.

We conducted a secondary data analysis using US Census ACS data provided by the City Health Dashboard (the Dashboard). The Dashboard includes all US cities with a 2017 ACS population of 50,000 or more (n=753), including three slightly smaller cities to ensure all states were represented (Burlington, Vermont; Cheyenne, Wyoming; and Charleston, West Virginia) and ten smaller New Jersey cities (population < 30,000). The complete list of 766 cities can be found on the Dashboard website ¹¹ and in Appendix Exhibit A1. For the purposes of this analysis, conducted at the city and census tract levels, broadband access is defined as the percentage of households with connections to high-speed broadband internet (including cable, fiber optic, and DSL connections).¹² Households with only cellular data plans are excluded from this metric because cellular plans do not support the range of internet services provided by high-speed broadband.¹³ We explored the trend from 2017-2019 for city and neighborhood broadband access.

We calculated the median percentage of households reporting broadband access and further disaggregated neighborhood-level estimates by median household income and neighborhood racial/ethnic composition. For median household income disaggregation we assigned cities and neighborhoods to income quartiles (low, low to medium, medium to high and high income) using ACS 2019 5-year estimates (ACS variable S1901_C01_012E) relative to the other cities and neighborhoods displayed on the Dashboard. For racial/ethnic composition disaggregation we categorized neighborhood racial/ethnic composition by whether a racial/ethnic group comprised over half of the total population in the city/neighborhood (ACS table: DP05). Race/ethnicity categories included Asian Americans (AA), non-Hispanic Black (Black), Hispanic/Latino (Hispanic), non-Hispanic White (White), American Indian, Native Hawaiian, and other Pacific Islander (AI or NH&PI, as a single group due to low counts). Neighborhoods without an absolute majority race/ethnicity were categorized as No Majority. In the ‘no majority’ group, the average distribution of White, non-Hispanic Black, Hispanic, and AA was 35%, 20%, 28% and 15%, respectively. As a sensitivity analysis, we also classified neighborhoods based on which racial/ethnic group had the highest population percentage, even if that percentage did not represent a majority of the city/neighborhood’s population. The results of this analysis are displayed in Appendix.

Across Dashboard cities, on average, 72% of households were connected to broadband internet in 2019. Income and racial/ethnicity were associated with broadband access. Median broadband access by median household income suggested that high income neighborhoods had the highest broadband access rate (85.3%), and low income neighborhoods had the lowest broadband access rate (51.4%). Median broadband access by race/ethnicity group suggested that AA majority neighborhoods had the highest broadband access rate (79.9%), followed by White majority neighborhoods (78.4%) and neighborhoods with no majority race/ethnicity (72.3%). Black and Hispanic majority neighborhoods have the lowest broadband access rate (51.5% and 59.7% respectively). At least 75% of the AA and White majority neighborhoods had higher broadband access rate than the top 25% of the Black and Hispanic neighborhoods. The upper quartile of broadband access among the Black and Hispanic majority neighborhoods were 62.7% and 69.0% respectively. In contrast, the lower quartile of broadband access among the AA and White majority neighborhoods were 71% and 70% respectively (Figure 1 & Table 1).

After stratifying by income, neighborhoods with majority non-white residents consistently have lower broadband access than White majority neighborhoods, including AA majority neighborhoods. Among low-income neighborhoods, the median broadband connection rate for AI and NH&PI majority neighborhoods was 42.9% (IQR: 36.1, 46.3), for Black majority neighborhoods was 45.7 % (IQR: 37.2, 53.5), for Hispanic majority neighborhoods was 48.8% (IQR: 41.1, 56.7), and for AA majority neighborhoods was 52.8% (IQR: 46.2, 60.0. These numbers were considerably lower than white majority neighborhoods (60.8%; IQR: 53.4, 67.5), and neighborhoods with no majority race/ethnicity (55.1%; IQR: 47.3, 62.4). Roughly similar racial/ethnic disparities were seen in high-income neighborhoods. 77.4% of households in AI and NH&PI majority neighborhoods, 76.8% in Black majority neighborhoods, 79.9 % in Hispanic majority neighborhoods had broadband access, while 85.9% of households in AA and White majority neighborhoods, and 83.8% of the households in neighborhoods with no majority race/ethnicity group were connected to broadband internet (Figure 2 & Table 1).

Between 2017 and 2019, access to broadband increased modestly, and disparities in broadband access diminished. Overall, the median broadband connection rate increased from 70% to 72% across all neighborhoods (Table 1). Broadband access increased 6.4% in low-income neighborhoods, a faster rate than the 1.3% increase rate in high-income neighborhoods. In each income stratum, increases in broadband access were larger in Black and Hispanic majority neighborhoods (2-7%) than in neighborhoods with other race/ethnic compositions. Broadband access in AA majority neighborhoods increased only slightly (0.7%-1.6%) over the 3-year period (Figure 3).

Figure 4 shows the geographical distribution of broadband access across the 766 Dashboard cities. Generally speaking, cities in the west had higher broadband access than cities in the northeast and south. Cities in the Great Lakes region and along the southern border had the lowest broadband access among the 766 included cities.

Four key findings emerge from our analysis of broadband access in US cities. First, in 2019, more than a quarter of households in the 766 largest US cities did not have broadband access at home. Second, households in low-income neighborhoods were less likely to have broadband access compared to households in high-income neighborhoods. Third, neighborhoods composed predominantly of minority households consistently had lower broadband access compared to White and no majority neighborhoods, regardless of income level. Fourth, although broadband access increased only modestly between 2017 and 2019, the increase was larger in low-income and minority predominant neighborhoods and had the effect of modestly reducing racial/ethnic and income disparities.

Despite progress made over the three-year period, our data indicate that substantial disparities persist in urban broadband access. Across the 766 cities analyzed, more households from low-income neighborhoods lacked access to high-speed internet than did households from higher income neighborhoods. Previous literature generally ascribes lack of broadband access to an absence of broadband infrastructure, unwillingness on the part of broadband providers to invest in such infrastructure, and the cost burden to individuals of paying for broadband service ¹⁴. However, it is unlikely that lack of broadband access in cities is driven by infrastructure-at least 30% of household have broadband access in most urban neighborhoods (see the lower quartile in Exhibit 4), suggesting broadband infrastructure is available in these neighborhoods. Instead, reducing the cost of broadband access, potentially through providing direct subsidies for broadband subscriptions and computer devices, could reduce disparities in broadband access.

However, our findings suggest that income alone cannot explain comparatively lower broadband connection rates in Black, Hispanic, and AI or NH&PI majority neighborhoods across income strata. If 70% of the households in a neighborhood were connected to broadband internet, this neighborhood would be more connected than 75% of Black and Hispanic majority neighborhoods, but less connected than 75% of AA and White majority neighborhoods. This suggests significant broadband disparities by racial and ethnic groups. Even among neighborhoods with high median household income, broadband access in White majority neighborhoods was 12% higher than in Black majority neighborhoods and 7.5% higher than in Hispanic majority neighborhoods. One explanation is differences in disposable income. Even within the same income quartile, Black and Hispanic households tend to have less family wealth such as savings and house ownership than White or Asian households, which may make Black and Hispanic households more sensitive to the cost of broadband subscription.¹⁵ In addition to affordability, previous studies also suggested that a number of factors, including education disparities, languages barriers, digital literacy, and home computer ownership in childhood, may contribute to the low broadband access among racial and ethnic minority groups regardless of their income stratum.¹⁶ In many cities, the disparity in broadband also likely reflects the legacy of structurally racist disinvestment and differential development practices. For example, neighborhoods redlined more than a half century ago demonstrate lower broadband access today,¹⁷ a reflection of the greater systematic social disparity in US cities. Investing in “soft” infrastructure, such as education and services, in urban settings with fewer non-white residents, may help to reduce racial/ethnic disparities in broadband access in urban areas, where broadband infrastructure is widely available.

We additionally observed a notable paradox in broadband access in AA neighborhoods. Median broadband access in low-income AA neighborhoods was considerably lower than in low-income White majority neighborhoods. However the majority of AA neighborhoods were in the high-income group, which had higher broadband access overall, thus driving up median broadband access rates for all AA neighborhoods and masking lower access specifically in the low-income stratum. This paradox further emphasizes the need for stratification in conducting social health research, especially among racial/ethnic groups that encompass a wide range of ethnic subgroups, such as AA and Hispanic populations.¹⁸

Broadband access is an emerging social determinant of health with impacts across the life course, affecting students’ ability to learn and adults’ ability to find jobs.¹⁹ The COVID-19 pandemic has substantially increased public understanding of the importance of broadband access. City policymakers should take advantage of recent infrastructure funding targeted at broadband access to eliminate disparities, not only in rural but in urban communities as well. With city-level maps of household broadband access for 766 large US cities, resources like the City Health Dashboard can be used by city policymakers to target investments and interventions to close the gap in broadband access.

Author contributions
Y.L. performed data analysis and wrote the manuscript. T.L. and P.H. conducted the initial metric research. B.R.S., L.E.T. and M.N.G. generated the research idea and supervised the whole work. All authors reviewed and contributed to the final version of the manuscript.
Data availability statement
Data are available on https://www.cityhealthdashboard.com/.
All experiments were performed in accordance with relevant guidelines and regulations

Benda, N. C., Veinot, T. C., Sieck, C. J. & Ancker, J. S. Broadband Internet Access Is a Social Determinant of Health! Am J Public Health110, 1123-1125, doi:10.2105/AJPH.2020.305784 (2020).
Graves, J. M., Abshire, D. A., Amiri, S. & Mackelprang, J. L. Disparities in Technology and Broadband Internet Access Across Rurality: Implications for Health and Education. Fam Community Health, doi:10.1097/FCH.0000000000000306 (2021).
Ortega, G. et al. Telemedicine, COVID-19, and disparities: Policy implications. Health Policy Technol9, 368-371, doi:10.1016/j.hlpt.2020.08.001 (2020).
Webb Hooper, M., Napoles, A. M. & Perez-Stable, E. J. COVID-19 and Racial/Ethnic Disparities. JAMA323, 2466-2467, doi:10.1001/jama.2020.8598 (2020).
LaRose, R., Gregg, J. L., Strover, S., Straubhaar, J. & Carpenter, S. Closing the rural broadband gap: Promoting adoption of the Internet in rural America. Telecommunications Policy31, 359-373, doi:https://doi.org/10.1016/j.telpol.2007.04.004 (2007).
MARTIN, M. Computer and Internet Use in the United States, 2018. (United States Census Bureau, 2021).
The White House. Fact Sheet: The Bipartisan Infrastructure Deal. (The White House,, 2021).
Prieger, J. E. & Hu, W.-M. The broadband digital divide and the nexus of race, competition, and quality. Information Economics and Policy20, 150-167, doi:https://doi.org/10.1016/j.infoecopol.2008.01.001 (2008).
Whitacre, B., Strover, S. & Gallardo, R. How much does broadband infrastructure matter? Decomposing the metro–non-metro adoption gap with the help of the National Broadband Map. Government Information Quarterly32, 261-269, doi:https://doi.org/10.1016/j.giq.2015.03.002 (2015).
Massey, D. S., Rothwell, J. & Domina, T. The Changing Bases of Segregation in the United States. The ANNALS of the American Academy of Political and Social Science626, 74-90, doi:10.1177/0002716209343558 (2009).
Dashboard, C. H. City Health Dashboard, <https://www.cityhealthdashboard.com/> (2021).
Dashboard, C. H. City Health Dashboard Technical Document, <https://www.cityhealthdashboard.com/technical-documentation> (2021).
Reisdorf, B. C., Fernandez, L., Hampton, K. N., Shin, I. & Dutton, W. Mobile Phones Will Not Eliminate Digital and Social Divides: How Variation in Internet Activities Mediates the Relationship Between Type of Internet Access and Local Social Capital in Detroit. Social Science Computer Review, 089443932090944 (2020).
Reddick, C. G., Enriquez, R., Harris, R. J. & Sharma, B. Determinants of broadband access and affordability: An analysis of a community survey on the digital divide. Cities106, 102904, doi:10.1016/j.cities.2020.102904 (2020).
Bhutta, N. C., Andrew C.; Dettling,Lisa J. ; and Hsu, Joanne W. . Disparities in Wealth by Race and Ethnicity in the 2019 Survey of Consumer Finances. (Board of Governors of the Federal Reserve System (U.S.), 2020).
Fairlie, R. W. Is There a Digital Divide? Ethnic and Racial Differences in Access to Technology and Possible Explanations (University of California, Latino Policy Institute and California Policy Research Center 2003).
County Health Rankings & Roadmaps. A digital disadvantage: Low broadband rates in urban, segregated pockets compound inequities (2022).
Spoer, B. R. et al. Neighborhood-level Asian American Populations, Social Determinants of Health, and Health Outcomes in 500 US Cities. Ethn Dis31, 433-444, doi:10.18865/ed.31.3.433 (2021).
Understanding the Broadband Access Gap, <https://www.huduser.gov/portal/pdredge/pdr_edge_featd_article_100614.html> (

Table 1: Broadband Access Rates by Income and Majority Race/Ethnicity, 2017-2019

Majority Race/Ethnicity	# of tracts	2017 Median (IQR)	2018 Median (IQR)	2019 Median(IQR)	% Increase 2017-2019
Low Income Tracts
AI or NH&PI	3	27.7 (26.5, 36.5)	33.5 (28.2, 40.0)	42.9 (36.1, 46.3)	54.9%
Black	2920	42.4 (34.4, 50.3)	43.8 (35.5, 51.8)	45.7 (37.2, 53.5)	7.8%
Hispanic	2078	45.3 (37.0, 53.9)	46.9 (38.6, 54.9)	48.8 (41.1, 56.7)	7.7%
AA	60	52.4 (47.0, 60.8)	53.0 (47.9, 60.1)	52.8 (46.2, 60.0)	0.9%
White	1744	58.3 (50.6, 65.7)	59.4 (51.8, 66.6)	60.8 (53.4, 67.5)	4.2%
No Majority	1501	52.3 (43.9, 59.5)	53.5 (45.6, 61.1)	55.1 (47.3, 62.4)	5.4%
Total	8306	48.3 (38.9, 57.3)	49.7 (40.4, 58.5)	51.4 (42.2, 60.1)	6.4%
Low to Medium Income Tracts
AI or NH&PI	1	59.4 (59.4, 59.4)	54.2 (54.2, 54.2)	63.2 (63.2, 63.2)	6.4%
Black	835	59.9 (53.5, 65.9)	61.3 (54.8, 67.3)	63.0 (56.5, 69.0)	5.2%
Hispanic	2055	58.2 (50.8, 64.7)	59.4 (52.4, 66.0)	61.4 (54.5, 67.6)	5.5%
AA	127	68.0 (61.8, 73.4)	68.3 (63.1, 73.6)	69.1 (61.9, 73.7)	1.6%
White	3552	68.0 (62.3, 73.4)	69.1 (63.5, 74.3)	70.3 (64.6, 75.4)	3.4%
No Majority	1737	65.5 (59.3, 71.7)	66.6 (60.6, 72.4)	68.0 (61.9, 73.5)	3.8%
Total	8307	64.5 (57.4, 70.9)	65.7 (58.7, 71.8)	67.0 (60.2, 73)	3.9%
Medium to High Income Tracts
AI or NH&PI	4	67.2 (64.0, 71.1)	65.6 (64.4, 68.4)	64.5 (64.1, 67.8)	-4.1%
Black	364	70.1 (64.1, 76.4)	71.3 (64.7, 76.5)	72.5 (66.0, 77.8)	3.5%
Hispanic	1096	69.0 (62.5, 75.0)	70.2 (64.2, 76.0)	71.7 (65.8, 76.8)	3.9%
AA	185	73.8 (69.2, 78.5)	74.4 (69.6, 78.0)	74.3 (68.4, 78.6)	0.7%
White	4891	76.3 (71.0, 81.0)	77.1 (72.1, 81.6)	77.8 (73.0, 82.3)	2.0%
No Majority	1765	75.0 (69.7, 80.0)	76.0 (71.0, 80.6)	76.6 (71.9, 81.0)	2.1%
Total	8305	74.9 (69.2, 80)	75.8 (70.4, 80.6)	76.6 (71.2, 81.2)	2.3%
High Income Tracts
AI or NH&PI	4	77.4 (72.6, 83.4)	79.6 (75.2, 83.0)	77.4 (73.8, 80.8)	0.0%
Black	135	75.0 (69.4, 80.6)	75.3 (68.4, 82.7)	76.8 (67.3, 82.7)	2.4%
Hispanic	243	78.7 (73.9, 83.5)	78.7 (74.4, 84.2)	79.9 (75.2, 85.1)	1.5%
AA	474	85.3 (80.7, 90.2)	85.8 (81.6, 90.1)	85.9 (80.8, 89.8)	0.7%
White	6148	84.7 (80.2, 88.5)	85.3 (81.0, 89.0)	85.9 (81.8, 89.4)	1.4%
No Majority	1304	82.9 (78.2, 87.6)	83.3 (78.9, 87.6)	83.8 (79.6, 88.1)	1.0%
Total	8308	84.2 (79.5, 88.3)	84.8 (80.2, 88.7)	85.3 (81, 89.1)	1.3%
Summary
All	33226	70.0 (56.5, 80.2)	71.0 (57.7, 81.0)	72.1 (59.1, 81.7)	3.0%
Notes: Low-income: ≤41,961 USD, low to medium income tracts: 41,961 – 60,052 USD, medium to high-income tracts: 60,052 – 84,681 USD, high income: ≥ 84,681 USD

No competing interests reported.

Appendix.docx

PDF

Reviews received at journal
16 Jun, 2022
Reviewers agreed at journal
13 Jun, 2022
Reviewers invited by journal
13 Jun, 2022
Editor assigned by journal
13 Jun, 2022
Editor invited by journal
10 Jun, 2022
Submission checks completed at journal
10 Jun, 2022
First submitted to journal
16 May, 2022

You are reading this latest preprint version

Racial/Ethnic and Income Disparities in Neighborhood-Level Broadband Access in 766 US Cities, 2017- 2019

Status:

Version 1

Abstract

Figures

Introduction

Study Data And Methods

Study Results

Discussion

Conclusion

Declarations

References

Table

Competing Interests

Supplementary Files

Status:

Version 1