Comparing broadband access to adoption in urban, suburban, and rural America
One of the very first posts I wrote on this site, in December 2020, was a comparison of how many people have access to broadband versus how many people have subscribed to broadband service. More than two years later, we have new FCC maps that measure broadband access, and new American Community Survey data that measures adoption. The view is fascinating. Only 64.4% of rural American households have access to broadband at 100/20 throughput. Most, 58.8%, subscribe to broadband, a gap of less than 6 percentage points. Even with new FCC maps, 98.5% of urban households have access to broadband, but only 73% subscribe. The number in the suburbs is only slightly better: 97% access and 76% adoption.
Before I go any deeper, I strongly recommend reading John Horrigan’s excellent analysis of this ACS data for Benton, and the McKinsey report using the same data. The former argues convincingly that the ACP program has succeeded in increasing adoption of broadband in America’s cities; the latter points out the persistent racial divide among broadband adoption and makes recommendations to close the gap.
There is much to be optimistic about in this data. 36% of rural American households don’t have access to 100/20 broadband. As we all know, help is on the way, in the form of $42.45 billion, plus other programs. The high adoption rate relative to access, 59% adoption against 64% access, suggests that the currently unserved and underserved are eager for access and will have a high take-rate.
On the other side, there is plenty of room to grow in urban areas. At 73% adoption, there are about 10.2 million homes that still don’t have broadband. That’s similar to the 12 million rural homes that don’t have broadband service currently. Another 17.4 million suburban homes don’t have broadband.
As I’ll detail more below in the methodological notes, this is 5-year averaged data from the American Community Survey, so it’s likely missing the more recent effects of the ACP program which have brought up adoption another few percentage points, from 72% nationally in the 5-year averaged ACS data vs 75.5 percent in the 1-year ACS data. This shows how critically important the ACP is: we’re improving urban broadband adoption, but we have more to do and don’t want to go backwards if it expires.
Ok some methodological notes. There isn’t a standard definition of geographies that includes urban, suburban, and rural. The Census has a rural and urban indicator, but it lacks suburban. In previous analysis, I’ve used a classification at the tract level from HUD. That hasn’t been updated for the 2020 Census so it would be very hard to work with. Instead, I’ve ordered block groups by their density and assigned the most dense block groups that make up 27% of housing units as urban. The least dense block groups that make up 21% of the housing units I’ve assigned to rural. And the middle 52% I’ve assigned to suburban. Those are the percentages that Americans self-identify as the classification of their neighborhoods.
For broadband access, I’ve used a threshold of 100/20 throughput from a wired technology or licensed fixed wireless (essentially the BEAD definition at 100/20). I’m done with 25/3.
For broadband adoption, I use the ACS’s answer to this question below, which essentially means wireline broadband, though it doesn’t ask the respondent for a throughput threshold. Previously, many analysts had misread the ACS Internet Use section and included cellular-only households as having broadband. I’m happy to say I haven’t seen much of that from the most recent ACS analysis.
The ACS 5-year data is the average for the responses to these questions from 2017-2021. I need to use the 5-year data because it goes down to the block group level.