Estimating Mortality Rates for the US Homeless Population

02/01/2024
Featured in print Digest

This figure is a line graph titled, Probability of Death Among US Homeless, Housed, and Housed Poor Population in a Six-Month Period. The y-axis labeled, mortality hazard rate of those ages 18 to 54 in 2020. It ranges from 0.0000 to 0.0125. The x-axis ranges from April 2016 to March 2022, showing the results for 6-month periods. There is a grey box from March 2020 until March 2022 that is labeled, COVID-19 pandemic.  The graph displays three lines representing different housing statuses: Homeless, Housed Poor, and Housed. From 2016 until 2019, all three lines exhibit a similar trend, remaining nearly horizontal. However, starting in late 2019 or early 2020, coinciding with the onset of the COVID-19 pandemic, these lines begin to diverge in terms of their rate of increase. The Homeless line experiences the most significant rise during this period, indicating a substantial increase. In contrast, the Housed Poor line shows only a slight increase, while the Housed line undergoes minimal change The source line reads, Source: Researchers’ calculations using data from the US Census Bureau.

While low socioeconomic status is well known to be associated with heightened mortality risk, little is known about the mortality rate of the most disadvantaged segment of the US population, people experiencing homelessness. In Life and Death at the Margins of Society: The Mortality of the US Homeless Population (NBER Working Paper 31843), Bruce D. MeyerAngela Wyse, and Ilina Logani illustrate the stark health disparities associated with homelessness.

The researchers follow 140,000 people recorded as sheltered and unsheltered homeless during the 2010 Census through 2022, tracking all-cause mortality using administrative data from the Social Security Administration (SSA). They compare homeless individuals’ mortality risk to that of a representative sample of the housed US population and examine heterogeneity by income, demographic characteristics, disability status, and the extent of observed family connections to identify especially vulnerable subsets of this already very deprived population.

Homeless individuals face 3.5 times the mortality rate of those who are housed, after accounting for differences in demographic characteristics and geography.

These estimates suggest that nonelderly homeless individuals face about 3.5 times the mortality risk of people who are housed, accounting for differences in demographic characteristics and geography. This disparity far exceeds the mortality gap between Black and White housed individuals (1.4) and between poor housed and all housed individuals (2.2). Mortality risk relative to the housed population differs over the lifecycle and is highest when these individuals are in their 40s and 50s, with a 40-year-old homeless person facing a similar mortality risk as a housed person nearly 20 years older.

Within the homeless population, employment, higher incomes, and more extensive observed family connections are associated with lower mortality. In contrast to the housed population, where White individuals have a lower mortality risk than those who are Black, homeless White individuals face about a 40 percent higher mortality rate than homeless Black individuals, a finding that may be related to the lower prevalence of substance abuse and behavioral health conditions in the latter group. These analyses also reveal the substantial health risks faced by people experiencing homelessness even when they are not sleeping on the streets, with those recorded in shelters in the census facing a similar mortality risk as those counted at unsheltered locations.

The researchers estimate that mortality in the homeless population rose by about one-third during the COVID-19 pandemic over the expected increase in mortality due to aging. While the proportional rise in mortality risk was similar for people who were housed, the pandemic appears to have affected a much larger share of the homeless population because of their substantially elevated baseline mortality risk.

—Leonardo Vasquez


The researchers acknowledge the Alfred P. Sloan Foundation, the Russell Sage Foundation, the Charles Koch Foundation, the Menard Family Foundation, the National Institute on Aging, and the American Enterprise Institute for their support of the Comprehensive Income Dataset project.