The Mortality of the U.S. Homeless Population

This paper provides the first national calculation of mortality, the most severe indicator of health and well-being, for the U.S. homeless population. We use a sample of 140,000 people who were sheltered or unsheltered homeless during the 2010 Census, by far the largest and closest to representative sample ever used to study the homeless population. These individuals, along with housed and housed poor comparison groups, are linked to Social Security Administration data on all-cause mortality from 2010-2022 to estimate the magnitude of health disparities associated with homelessness. We find that non-elderly people experiencing homelessness have 3.5 times higher mortality than those who are housed, accounting for differences in demographic characteristics and geography. A 40-year-old homeless person faces a similar mortality risk to a housed person nearly twenty years older and a poor housed person nearly ten years older. The mortality rate of the homeless population relative to the housed is highest when individuals are in their 30s and 40s but falls in relative terms starting around age 50. Within the homeless population, people who are Black, female, and Hispanic have lower relative mortality risk than their white, male, and non-Hispanic counterparts. Employment, higher income, and observed family connections are also associated with lower mortality, but mortality risk is similar for people who were observed as sheltered and unsheltered homeless in 2010. Homeless individuals’ mortality rose by 33 percent during the COVID-19 pandemic, an increase that, while similar in proportional terms to the increase for the housed population, affected a much larger share of the homeless population due to their substantially elevated baseline mortality rate. These findings elucidate the persistent hardships associated with homelessness while also identifying more vulnerable segments of an already exceptionally deprived population.