Diversity in a region’s housing characteristics can be a strong indicator of diversity in its population: a healthy variation in local residences’ size, age, and cost signals accessibility to a broader range of social and economic groups. Policymakers currently rely on federal data sources like the American Community Survey to assess housing diversity, but this information may provide an incomplete picture. In these records, factors like housing value are averaged across sizable census tracts, obscuring some of the significant differences that may exist within a region.
The U.S. Census Bureau has partnered with SDAL to develop more fine-tuned measurements of housing diversity using publicly available data sources. These new methods could tell decision-makers more about the specific mix of housing stock that characterizes a region, providing valuable indicators of regional diversity and economic health.
Typically, diversity metrics are developed using data from federal surveys such as the American Community Survey. The most detailed geographic unit of measurement used in these surveys is at the U.S. Census tract, a sizable region that encompasses between 1,200 and 8,000 residents. More precise measures are available at the housing unit level, however. These are maintained through publically available, county and city real estate assessments.
Left: American Community Survey (ACS) 2009-2013 5-year estimates of median house value for owner occupied housing units by tract for Arlington County, VA.
Right: Arlington County, Virginia Real Estate Assessment (REA) 2013 values of housing units in 3 census tracts circled in panel on the left. The higher level of granularity of the data allows the heterogeneity within a tract to be presented and compared to surrounding tracts.
Using Arlington’s 2013 real estate assessment data, SDAL researchers successfully computed Simpson diversity indices (Simpson 1949) accounting for housing unit value, year built, and property type (Molfino et al. 2017.) The precise locations of residential homes included in these data sets provide a greater degree of flexibility for assessing diversity within census tracts. This allows reviewers to observe data at the census block-group level or even within smaller neighborhood groupings.
Our findings demonstrate that there is substantial value in developing more geographically refined diversity measures using local data. The distributions in the diversity indexes illustrated here show that there are significant differences within regions that are not captured when observations are limited to the census tract level. Locally collected data could be an indispensible tool for targeting policies related to housing diversity toward specific neighborhoods, including measures to increase the availability of affordable housing or access to public services.
Maps of Simpson indices of housing value by census tract (left) and by block group (right.) Simpson indices for census tracts or block groups with fewer than 25 observations were excluded (white regions.) (Source: Arlington County Real Estate Assessment Data 2013.)
In the context of our communities where we live, learn, work, and play, diversity may reflect a variety of factors that impact residents’ quality of life, such as housing prices, infrastructural resilience, and safety. Identifying the characteristics of these neighborhoods in more detail will provide more useful insights to government leaders and policymakers.