51 and 61 years old grew from 20 to 26% (Henry, Bishop, et al., 2018).
The story for families with children is different. In the New York City data, from 1988 to 2005, the modal age of heads of families remained 21–23 (Culhane et al., 2013). More recent national data on families suggest an older typical age, as only 23% of adults in families using shelters at some time during 2017 were between the ages of 18 and 23. There was only a very small aging trend between 2007 and 2017 (Henry, Bishop, et al., 2018). Because infants and preschool children are common in families experiencing homelessness, it may be the age of children rather than the age of their parents that is critical for families. About half of the children in families using shelters over the course of a year are under the age of 6, and 11% are infants less than 1 year old (Henry, Bishop, et al., 2018). The slight increase in age of mothers may be because the mean age of U.S. mothers generally at the time their first child was born increased from 24.1 in 1988 to 26.8 in 2017 (Martin et al., 2018; Mathews & Hamilton, 2002). The number of adult men in homeless families is increasing (Henry, Bishop, et al., 2018), but this may be a function of the increasing willingness of shelters to accommodate couples rather than the changing composition of the families themselves.
The last decade has seen important changes in the numbers of people experiencing homelessness—changes that are particularly marked for people with chronic patterns of homelessness and for veterans. Overall, the number of people counted nationally in PIT counts decreased almost 15% from about 647,000 in 2007 to about 553,000 in 2018. Most of the drop of close to 100,000 total people was among people who were unsheltered. The numbers of people in shelter (on a single night) stood at about 391,000 in 2007, rose a bit during the aftermath of the Great Recession, came back to 391,000 in 2015, and has dropped a bit—to 358,000—since then. The additional progress in the last 3 years has been partially offset by a smaller increase in the unsheltered population (Henry, Mahathey, et al., 2018).
The number of people with chronic patterns found on the night of the counts decreased 26% from about 120,000 to less than 89,000 over the same period (Henry, Mahathey, et al., 2018).15 We think that decrease is real, not just a matter of reporting, because we do not think that systematic changes have taken place in the way communities nationwide count chronically homeless individuals. We also think the drop is consistent with the success of permanent supportive housing programs that combine housing with voluntary services, as we will discuss in Chapter 3, on ending homelessness for people who experience it.
The number of veterans experiencing homelessness nationally fell even more dramatically from about 73,000 in 2010 (the first year data were available), to about 38,000 in 2018 (Henry, Mahathey, et al., 2018). We believe that federal and local policy efforts that we will describe later, in Chapters 3 and 4, account for cutting veteran homelessness nearly in half. These decreases, especially for veterans, show what can be done with concerted effort.
Total Numbers Over a Day, a Year, or a Lifetime
We have HUD's estimate of 553,000 people on a single night in January, 2018. Given the likely undercounts of people staying in unsheltered locations, the total number of people experiencing homelessness in the U.S. nationwide on a particular night in January 2018 was probably more than 600,000.
But what about longer periods? Over the course of a year, we believe the number is closer to 1.6 million people. HUD's estimate of the number of different people who used a shelter at some point during the course of a year was just over 1.4 million in 2017, dropping from almost 1.6 million in 2007.16 To get a total that includes people who were unsheltered, we might add another 200,000. That is the unsheltered PIT count, which as we have already pointed out, misses many people who were on the streets that night. It also misses people who were unsheltered on a different night during the year. But many people who experience unsheltered homelessness also use shelters at other times and are already in the 1.4 million. If about the same number who are missed also use shelters at any point over the year, then the total number of people experiencing literal homelessness over the course of a year may be 1.6 million.17,18
Thus far we have estimates of people homeless on a particular night (sheltered and unsheltered, about 600,000 people) and people homeless over the course of a year in shelters (about 1.6 million people). But a year is a short period of time. Much larger estimates of the number of people experiencing homelessness come from asking people who are living in conventional housing about homelessness they experienced at some time during their lives and not just in the past year or two.
Back in 1990, researchers led by Bruce Link undertook a study of public attitudes toward homelessness with a rigorous sample survey of adults in households with telephones in the continental United States. Because they thought that those attitudes might be influenced by people's personal experiences of homelessness, they decided to ask about those experiences and then to ask some follow‐up questions to anyone who acknowledged having been homeless in the past. Survey organizations charge researchers by the minute for asking questions, thus the organization Link hired to do the survey had to come up with some sort of estimate of the number of people who would be asked the follow‐up questions. Organizational representatives decided it would be such a small number that it would not be worth any charge. (After all, this was a household survey—anyone currently homeless or staying in a prison or mental hospital—who might be at higher risk of having been homeless in the past—would not be included.) They guessed wrong.
Fully 14.0% of the 1,507 survey respondents answered yes to the question “Have you ever had a time in your life when you considered yourself homeless?” and 4.6% said they had been homeless between 1985 and 1990 (Link et al., 1994, p. 1909). Respondents who classified themselves as having been homeless were asked three follow‐up questions: “While you were homeless, did you ever (1) sleep in a park, in an abandoned building, in the street, or in a train or bus station?; (2) sleep in a shelter for homeless people or in another temporary residence because you did not have a place to stay?; (3) sleep in a friend's or relative's home because you were homeless?” (p. 1909). A little over half of the people who said they had been homeless at some time in their life, 7.4% of the survey respondents, said yes to one of the first two questions and were classified as literally homeless.19 The other people, who said yes to only to the last question, were classified as precariously housed. The sequence of questions did not allow the researchers to determine definitively the percentage of people who had been literally homeless in the past 5 years, but 3.1% of the sample said they had been homeless in the past 5 years and also met the criteria for literal homelessness at some point in their lives.
The researchers asked about total duration of homelessness (including doubling up): the most frequent reply (46%) was between a month and a year. Only 8% had been homeless for less than a week, 33% between a week and a month, and 13% for over a year. In 1990 there were slightly over 185 million adults living in the United States, so the researchers estimated that 13.5 million adults had been literally homeless at some time in their lives, and nearly 26 million had considered themselves homeless if doubling up is included (p. 1910). The fact that many people are homeless for fairly short periods is also shown in later analyses of shelter records by Culhane et al. (2007; Kuhn & Culhane, 1998) and explains why the numbers for lifetime homelessness are so much larger than the numbers for any given night or year.
Estimates of people homeless or precariously housed at some time during a year show a substantial increase between the late 1980s and the early 2000s. The percentage of adult respondents to the national General Social Survey who replied yes to a combined question about whether the respondent “had to temporarily live with others or in a shelter or ‘on the