Urban. Suburban. Rural. How Do Households Describe Where They Live?
Shawn Bucholtz, Director of PD&R's Housing and Demographic Analysis Division.
New 2017 American Housing Survey neighborhood description data and data products shed light on how "suburban" America really is. A recent webinar explores this topic in more detail. The webinar featured experts from HUD, the U.S. Census Bureau, and Indeed.
Although most existing federal definitions of “urban” and “rural” do not include a “suburban” category, data from HUD and the U.S. Census Bureau (Census) confirm what researchers have long believed: most Americans live in the suburbs. According to data HUD and Census collected in the 2017 American Housing Survey (AHS), 52 percent of U.S. households describe their neighborhood as suburban, 27 percent describe their neighborhood as urban, and 21 percent describe their neighborhood as rural. The 2017 AHS data also show that federal definitions accurately distinguish urban neighborhoods from rural areas while underscoring the need for an official definition of suburban.
It’s a fact: Most of America is suburban
The 2017 AHS included a question asking respondents if they would describe their neighborhood as urban, suburban, or rural. HUD wanted to replicate the results of a 2015 survey conducted by economist Jed Kolko and colleagues at the online real estate company Trulia, which asked 2,000 people the same question. Drawing from a sample size of more than 55,000 households, the AHS results were nearly identical to Trulia's. Similar research from the Pew Research Center confirmed that most households in the United States describe their neighborhood as suburban. "With multiple national surveys reaching the same conclusion, the notion that the majority of Americans live in the suburbs is no longer an anecdote — it is a fact," says Shawn Bucholtz, head statistical officer and director of housing and demographic analysis at HUD.
Kolko, now a chief economist at Indeed, collaborated with Bucholtz on the AHS neighborhood description question. “There was a research purpose for understanding different housing dynamics, but more broadly, there’s long been confusion when you talk to people about what urban, suburban, and rural mean,” Kolko explains. “It is critical to get these classifications right as they play an important role in how HUD and other federal agencies allocate billions in tax dollars in communities,” Bucholtz adds.
How do current urban/rural definitions measure up?
When AHS respondents' answers to the neighborhood description question are compared with how their addresses are classified by the two main federal definitions of urbanization, the results are mixed.
The AHS data show that, when distinguishing urban from rural areas, definitions of urbanization from both the Census Urbanized Areas and the Office of Management and Budget's (OMB’s) metropolitan and micropolitan statistical areas closely match respondents' perceptions. Ninety-five percent of households living in Census Urbanized Areas (urban areas with more than 50,000 people) describe their neighborhood as either urban or suburban, and 80 percent of households in areas that Census defines as rural describe their neighborhood as rural. Likewise, the lion's share of households living in OMB's metropolitan statistical areas (86 percent) describe their neighborhood as urban or suburban, whereas a large percentage of households (72 percent) living outside of metropolitan or micropolitan statistical areas describe their neighborhood as rural.
At the same time, the 2017 AHS data reveal that existing definitions obscure the fact that most Americans live in suburbs. The 2017 AHS data show that Census Urbanized Areas and OMB metropolitan statistical areas, as currently defined, are mostly suburban. Nearly twice as many households in Census Urbanized Areas describe their neighborhood as suburban (63 percent) than urban (32 percent). Roughly twice as many households in OMB metropolitan statistical areas consider their neighborhood to be suburban (57 percent) compared with urban (29 percent).
AHS neighborhood description data show that even central cities — which are presumed to be the most urban part of metropolitan areas — are quite suburban. A slight majority of households (51 percent) living within the central city of a metropolitan area describe their neighborhood as urban, whereas nearly half (47 percent) describe their neighborhood as suburban. For areas outside of central cities but within a metropolitan area, most respondents (64 percent) describe their neighborhood as suburban.
Table 1: AHS Neighborhood Description Results: Central/Outside Central Cities
|Outside Central Cities||14%||64%||22%|
According to Kolko, "Often, when researchers attempt to compare cities and their suburbs based on existing publicly available data, they rely on this central city versus outside central city distinction as the way of separating urban from suburban. This new data show how much you actually miss by doing that."
When it comes to individual Urbanized Areas or metropolitan statistical areas — including Los Angeles, Chicago, Dallas-Fort Worth, Houston, and others—the data tell the same story: these areas, with few exceptions, are majority suburban. In fact, all the large metropolitan areas that the AHS surveys are majority suburban.
Table 2: AHS Neighborhood Description Results: Specific Metropolitan Areas
|New York City||47%||49%||4%|
A better way to classify America’s neighborhoods
To facilitate these comparisons and make it easier to accurately classify neighborhoods based on the descriptions of those who live there, HUD created two new data products using the AHS neighborhood description data. The American Housing Survey Neighborhood Study Summary Tables Workbook — a Microsoft Excel workbook containing 24 data tables — enables users to compare various definitions of urban, suburban, and rural to survey respondents' descriptions of their neighborhoods.
Motivated by Trulia's and Pew's previous work, HUD also used the AHS neighborhood description data to create a nationwide small area urbanization classification product based on people's descriptions of their neighborhood called the Urbanization Perceptions Small Area Index (UPSAI). UPSAI uses a machine-learning algorithm to classify each census tract as suburban, urban, or rural.
Census survey statistician Emily Molfino collaborated with Bucholtz and Kolko on the UPSAI. "Our main motivation was that we wanted to classify small areas as urban, suburban, or rural. We noticed there was a big demand for this type of data product, and we saw an opportunity to build upon what had done before with the American Housing Survey’s larger sample size. We wanted to build the next-generation product," Molfino says.
To create HUD's UPSAI product, analysts first applied machine-learning techniques to the AHS neighborhood description question to build a model that predicts how out-of-sample households would describe their neighborhood (urban, suburban, or rural) given regional and neighborhood characteristics. Analysts then applied the model to the American Community Survey tract-level aggregate regional and neighborhood measures, creating a predicted likelihood that the average household in a census tract would describe their neighborhood as urban, suburban, or rural.
Table 3: UPSAI tract-level file
HUD's UPSAI makes it easy to roll up the tract classifications into county and higher-level geographic averages, which is helpful because a considerable amount of critical data is available at the county level but not at the tract level. For example, with the UPSAI, users can combine the monthly unemployment rates from the U.S. Bureau of Labor Statistics with estimates of how urban, suburban, or rural counties are, on average, to see what's going on in urban, suburban, and rural unemployment.
Molfino adds, "the UPSAI format is very flexible in that it allows users to aggregate our product to any other level of geographic areas such as counties or congressional districts. I am very excited about how the final data product is easily accessible to data users no matter their experience."
Looking toward the future
Right now, federal experts are planning for the next versions of the Census Bureau Urbanized Areas, metropolitan statistical areas, and other urban/rural definitions using the 2020 decennial census data. "Official discussions about the next generation of definitions are already underway. We wanted to contribute to those discussions by providing data showing how well those definitions align with people's descriptions of their neighborhood," says Bucholtz.
For more information, see the webinar, “How Do Households Describe Where They Live?”