In the Midst of Plenty. Marybeth Shinn

suggest this number would be still higher—perhaps 27%—if family shelters did not exclude them.

Figure 1.1 Percentage of U.S. Population Using Shelter Over the Course of a Year (2016–2017) by Age Group. Numbers of people experiencing homelessness from Supporting Resources for the 2017 Annual Homeless Assessment Report to Congress Part 2 (Henry, Bishop, et al., 2018). Numbers of people in the U.S. population for 2017 from annual estimates of the resident population by single year of age and sex for the United States, States, and Puerto Rico Commonwealth: April 1, 2010 to July 1, 2017 (U.S. Census Bureau, 2018).

HUD does not include people, overwhelmingly women, who use domestic violence shelters, in these estimates, but the number of beds in domestic violence shelters is not large, so this does not account for the difference (Henry, Bishop, et al., 2018). Women, especially women with children, may be more able than men to persuade kin or non‐kin to take them in to doubled‐up households,¹³ and they have access to some additional safety‐net resources. Unaccompanied young people (children and youth under the age of 25) are somewhat more likely to be male, especially if they are on the street, but unaccompanied youth include more women and girls when compared with all adults who are homeless on their own (Henry, Mahathey, et al., 2018; Toro, Dworsky, et al., 2007).

African Americans and Native Americans are especially likely to become homeless (Burt et al., 1999; Hopper & Milburn, 1996). African Americans are particularly heavily represented among families who use shelters, with 52% of that population identifying as black or African American in 2017. By comparison, the African American share of the family population with incomes below the federal poverty level was 23% in the same year. Both shares have dropped somewhat in recent years (Henry, Bishop, et al., 2018).

Homelessness is largely an urban phenomenon. People who experience homelessness over the course of a year are more likely to be in principal cities¹⁴ (72.5%) than either the entire U.S. population (32.5%) or people living in poverty (39.6%), although the number in suburban and rural locations has increased from 23.1 to 27.5% since 2007 (Henry, Bishop, et al., 2018). Here again, homeless services may shape or distort our understanding of people who experience homelessness. Most services are located in cities, so people who become homeless in nearby areas without shelters may migrate there. A study that used HMIS data to look at migration patterns in two states where that was possible, Iowa and Michigan, documented some migration from suburban areas into cities, based on the zip code of the last permanent address before the person entered a shelter. The study found essentially no migration out of rural areas (Leopold, Culhane, & Khadduri, 2017).

      Youth who become homeless may or may not differ from the general population in their locale with respect to race or ethnicity (Toro, Dworsky, et al., 2007), but they are disproportionately sexual minorities. Estimates of the proportion of lesbian, gay, bisexual, and transgender (LGBT) youth range from 6 to 40% (Durso & Gates, 2012; Toro, Dworsky, et al., 2007). The 40% estimate comes from a national survey of 354 organizations serving homeless youth (primarily 18 and older) that made special efforts to include agencies serving LGBT youth (Durso & Gates, 2012); the midpoint of earlier estimates is closer to 20% (Toro, Dworsky, et al., 2007). The recent Voices of Youth Count Survey conducted in 2016 and 2017 found no significant difference between urban and rural areas in the prevalence of homelessness for young adults 18–25 (Morton et al., 2018), but that survey also was relatively small and included couch surfing as well as literal homelessness.

Changes Over Time

The patterns seen today by age, gender, race, and family status are a relatively recent phenomenon. After the “Hoovervilles” of the Great Depression (shanty towns of people without jobs) and up until the early 1970s, most people who experienced homelessness were older white men who no longer were able to do the physically demanding work of their younger years, currently had no job or intermittent, poorly paid jobs, and often had disabilities (Bogue, 1963). These men rarely slept on the street, although they were certainly “inadequately housed” in single‐room occupancy hotels (with no kitchen and a bathroom down the hall), mission dormitories, or flophouses. Many of the latter were divided into windowless five by seven‐foot cubicles with partitions that did not extend to the ceiling or floor. The wire mesh that filled the gap, allowing for minimal privacy and security, gave rise to the moniker “cage hotels” (Rossi, 1989, p. 30). Before public drunkenness was decriminalized, some slept in police stations and jails. Researchers focused more on the men's lack of social ties than on their housing circumstances (Bahr, 1973; Grigsby, Baumann, Gregorich, & Roberts‐Gray, 1990). Although several studies estimated that only about a quarter of men were alcoholics, researchers emphasized this problem, for example, entitling a book about New York's skid row Old Men Drunk and Sober (Bahr & Caplow, 1973). Many observers thought that the problem of homelessness would disappear as this older generation of men came to the end of their lives (Bogue, 1963).

      But homelessness did not die off. It changed. By the late 1970s, “the new homeless,” younger men, often African American, along with some women and even families, began to emerge (New York City Commission on the Homeless, 1992). Further, the shrinkage of skid rows associated with urban renewal made the residual “old” homelessness more visible. The decriminalization of vagrancy and public drunkenness meant that people who might once have sobered up in jail were now on the streets (Shlay & Rossi, 1992). Efforts to count and categorize people experiencing homelessness led to wildly differing estimates of the composition and characteristics of people experiencing homelessness during the 1980s. Nevertheless, across 60 studies conducted from 1981 to 1988, 26% of people identified as homeless were women, over 40% were black, and the median age was 37 (Shlay & Rossi, 1992), a far cry from the older white men of the 1960s.

      The age distribution of homelessness has continued to change, quite separately for single men and for parents in families. Men born in the latter half of the baby boom, from the mid‐1950s to the mid‐1960s, continued to dominate the numbers from 1988 to 2010, both in decennial census data (when the census conducted shelter counts) and in data from the Department of Homeless Services in New York City, which has the longest, most complete records of shelter usage. The age at which a single man was at highest risk of being found in a shelter (relative to the numbers in the overall population), peaked at 34–36 in 1990, 37–42 in 2000, and 49–51 in 2010 (Culhane, Metraux, Byrne, Stino, & Bainbridge, 2013).

      In the New York data, it is not the same people who continue to experience homelessness across the different decades, but newcomers to the homeless system are most often drawn from the same cohort of late baby‐boomers (Culhane et al., 2013). In‐depth interviews with homeless individuals sampled at a New York City drop‐in center for older adults found that roughly half of the sample (42 of 79) had led conventional lives with long periods of residential stability and employment through middle age (e.g., a grocery store manager, an army colonel, a fundraiser for a nonprofit) until some event, and usually a cascade of events, pitched them into homelessness. The other 37 people had more long‐standing patterns of housing instability, although not necessarily literal homelessness (Shinn et al., 2007).

Similarly, in San Francisco, successive groups of literally homeless adults found at shelters and free meal programs in San Francisco in each of four time periods, had a median age of 37 in 1990–1994 and a median of 46 in 2003. The earliest group had been homeless for 2 years on average, but the last group reported 6 years of literal homelessness (Hahn, Kushel, Bangsberg, Riley, & Moss, 2006).

      The national data reported by HUD show a similar aging trend. Among people experiencing homelessness as individuals, the percentage 62 years and older grew from 4% in 2007 to 8% in