What Does It Mean to Be Homeless?
How Definitions Affect Homelessness Policy
Andrew Sullivan (University of Central Florida)
Organizations and policymakers have recently brought the definition of homelessness to the forefront, including multiple reports by the Government Accountability Office, and the Homeless Children and Youth Act of 2021 which was introduced but not passed. Much of this debate stems from the U.S. Department of Housing and Urban Development (HUD) not including households living doubled-up – i.e., sharing housing due to economic hardship, loss of housing, or a similar reason – as homeless, contrary to many other definitions of homelessness. As such, reviewing how federal agencies measure homelessness provides insight into how they define the problem of housing insecurity and what impacts measurement has on determining where services are most needed. How scholars and policymakers define homelessness alters its measurement and perceived prevalence, driving where agencies target assistance and shaping policy solutions.
In a new article in Urban Affairs Review “What does it mean to be homeless? How definitions affect homelessness policy,” I show how the definition of homelessness changes the understanding of it and affects which places experience high rates of homelessness, leading to disparate impacts in types of communities receiving resources. For example, only 29% of the 20% of school districts with the most unsheltered and sheltered homelessness are also in the 20% of districts with the most doubled-up students or students living in hotels/motels. School districts with high rates of homelessness when including doubled-up students tend to be more rural, have a higher poverty rate, and have a larger share of students identify as Black or Hispanic compared to districts with a high rate excluding these students. These findings suggest what types of communities would receive more resources from expanding HUD’s definition to include households doubling up such as through the Homeless Children and Youth Act.
Conceptually, using a different definition could result in no change in determining where the problem is severe if rates of homelessness are proportional by definition; communities with high rates of people living in homeless shelters may also have high rates of doubling-up. However, this is an empirical question which past research has yet to explore. As I find communities differ in rates of homelessness depending on the definition used and aspects such as demographics, urbanicity, and poverty, the definition leads to disparate impacts, meaning it disadvantages certain groups in favor of others.
I first reviewed four definitions of homelessness from United States’ federal departments: Housing and Urban Development (HUD), Education (ED), Health and Human Services (HHS), and Agriculture (Food and Nutrition Act/FNA). These definitions vary due to the context, reason, and length of the living situation, presented in Table 1 below. For example, HUD considers someone as homeless solely based on context, such as living unsheltered or in emergency shelter. ED includes additional contexts, primarily sharing housing, but only if the reason for sharing housing is involuntary. Alternatively, the FNA only considers context and length, as people can receive benefits if sharing housing regardless as reason, but only if in the situation for 90 days or fewer.
Table 1. Homeless Status by Definition and Housing Status
Notes: The above table shows whether a hypothetical household with a child or youth would be considered homeless by each of the four federal agencies/acts. Involuntarily sharing housing means sharing housing with other persons as a result of loss of housing, economic hardship, lack of a safe alternative, or a similar reason. Voluntarily sharing housing means sharing housing with other persons not from loss of housing, economic hardship, or a similar reason. Additionally, involuntarily implies an intention by the household of only temporarily sharing housing.
Building from HUD’s and ED’s definitions, I then find that communities with high rates of homelessness not only differ depending on the definition used, but that they differ in key characteristics including demographics and urbanicity. To find this, I constructed estimates of student homelessness under a broad and a strict definition using data from ED. I then compared the 20% of school districts with the most homelessness under HUD’s definition to the 20% of districts with the most under ED’s definition. As residential context also influences which educational supports a student experiencing homelessness needs, understanding the form of housing insecurity students in a district face can help administrators provide the best and appropriate supports. Two districts with high rates of homelessness both have students in need and that require support, but what the support looks like substantially differs if one district has many students living doubled-up and the other many living in shelters.
I also show how characteristics of a district relate to larger discrepancies between definitions. A larger discrepancy implies communities have a large gap in homelessness between the definitions. A change in the definition would have the largest impact on these communities. I found that that communities that are suburban or a town with a larger share of the population being racial minorities could have the most to gain from expanding HUD’s definition.
Last, given the recent increase in scholarly work on aggregate-level studies of homelessness across fields – economics, education, public policy, and sociology, among others– clearly defining who counts as homeless in research must occur to obtain an understanding of the causes of homelessness and which interventions help. Doing so also makes use of the most variation, as policies may affect people with certain characteristics but not others such as children and youth doubled-up.
Andrew Sullivan is an assistant professor in the School of Public Administration at the University of Central Florida. His research focuses on collaborative governance, homelessness, and Continuums of Care.