Homelessness: Programs and the People They Serve | Findings of the National Survey of Homeless Assistance Providers and Clients

On most basic demographic characteristics, homeless clients differ considerably from the population of the United States. In addition, the parents in homeless families using services, who comprise 15 percent of the homeless clients to NSHAPC survey, also differ from single homeless clients on many of these same factors. Table 2.1 presents the relevant information. The Ns shown at the top of this and all other tables are unweighted. Percentages are based on weighted data.

Demographic characteristics of all U.S. adults appear in the first column of table 2.1 as a point of comparison to the homeless clients who participated in the NSHAPC survey. The second column describes these homeless clients. The third and fourth columns break out NSHAPC homeless clients into those who are with their own children (homeless families), and those who are not (single homeless clients).

Sex

As in other studies, men dominate among homeless clients, comprising 68 percent of this group compared to 48 percent of all U.S. clients in 1996 (figure 2.1).⁵ However, these statistics mask considerable differences in the sex of homeless clients in homeless families and single homeless clients. Among homeless clients in families, 84 percent are women and 16 percent are men (figure 2.2). Among single homeless clients, the sex ratio is reversed; only 23 percent are women and 77 percent are men (table 2.1).

Race/Ethnicity

Homeless clients are about equally divided between non-Hispanic whites and blacks (41 and 40 percent, respectively), with 11 percent Hispanics, 8 percent Native Americans, and 1 percent "other" (figure 2.3). Compared with all U.S. adults in 1996, homeless clients are disproportionately black non-Hispanics (11 versus 40 percent) and Native American (1 versus 8 percent).⁶ The racial/ ethnic makeup of homeless clients does not differ by family status.

Age

The overall statistics on the ages of homeless clients mask very great differences between homeless clients in families and single homeless clients. Clients in families are much younger, as shown in their higher probability of being ages 24 and younger (26 versus 10 percent) and lower probability of being ages 55 and older (less than 0.5 percent versus 9 percent) (figure 2.4). This age distribution is quite different from that of all U.S. adults, in which only 14 percent are under 25 years of age and 28 percent are ages 55 and older.⁷

Marital Status

Forty-eight percent of homeless clients have never married. Among the 52 percent who have been married at one time or another, most have seen those marriages dissolve through divorce (24 percent) or separation (15 percent) without subsequently entering into another marriage. These patterns are repeated among single homeless clients, and modified somewhat among homeless clients in families (figure 2.5). The latter group is more likely to report being married (23 versus 7 percent) and less likely to report being divorced (13 versus 26 percent). In addition, they are less likely never to have married (41 versus 50 percent), but the proportion who have never married is still high compared to all U.S. adults at 23 percent.⁸

Education

Thirty-eight percent of homeless clients have dropped out of high school, while for 34 percent, a high school diploma is their highest level of completed education. Fewer, but still more than one-quarter, have some education beyond high school. These figures differ considerably between clients in families and single homeless clients (figure 2.6). Clients in families are more likely to have ended their education before completing high school (53 versus 37 percent), and less likely to have exactly a high school diploma or G.E.D. (21 versus 36 percent). However, the two groups are equally likely to have received some education beyond high school. Homeless clients are less educated than the adult U.S. population: only 25 percent of American adults (those ages 25 and older) have less than a high school education, 34 percent have a high school diploma, and 45 percent have some education beyond high school.⁹

Veteran Status

Twenty-three percent of homeless clients are veterans. Examined separately by sex, 1 percent of homeless women are veterans compared to 33 percent of homeless men. For men, this proportion is not different from the 31 percent of the general client male population whom the Department of Veterans Affairs estimates were veterans in 1996, but somewhat lower than the 40 percent of veterans among homeless men found in a systematic synthesis of data from other studies of homeless populations (Rosenheck et al. 1996).

The Children of Homeless Clients

Many more homeless clients are parents than is indicated by the proportion who have at least one of their children with them. Among homeless women, 60 percent have children under age 18, but only 65 percent of them live with at least one of these children. Among homeless men, 41 percent have children under age 18, but only 7 percent of these fathers live with at least one of their own children.¹⁰ Looked at from the children's perspective, 28 percent of minor children of homeless parents live with that parent, while 72 percent do not.

Homeless families have, on average, two children. Members of these families comprise 34 percent of all homeless people using services. As this figure (34 percent) is quite different from the figure noted earlier—15 percent of homeless clients are heads of homeless families—some clarification is in order. The two figures illuminate a common confusion about the term "homeless family." Only 15 percent of homeless households contain an client and at least one minor child, which is a common definition of "family" used in the context of homelessness research. However, when one counts clients and children together, 34 percent are in families. Two-thirds of these are children.

Children in homeless families using services are fairly evenly divided between males (53 percent) and females (47 percent), which does not differ from American children generally (table 2.2). They are disproportionately younger than school age (ages 0 to 5) compared to all U.S. children (42 versus 34 percent).¹¹ Parents report that almost half (45 percent) of these children ages 3 to 5 attend preschool. In addition, almost all (93 per- cent) of school-age children (ages 6 to 17) are reported to attend school regularly.¹²

Homeless clients in families (table 2.1) and the children themselves (table 2.2) are similarly distributed among racial/ethnic groups. These similarities are due in large part to the fact that children's race/ethnicity was attributed from that of their parents, but also implies that the number of children homeless with their parent(s) does not differ systematically in relation to the parent's race or ethnicity.

NSHAPC parents' reports of their children's school attendance can be compared with data about school enrollment of U.S. children (table 2.2). Forty-five percent of the 3- to 5-year-old children accompanying homeless NSHAPC clients are reported to be attending preschool. Parents also say that 93 percent of their children ages 6 to 17 attend school regularly. The closest comparable figures for all U.S. children are for enrollment rather than for attendance. They indicate that 49 percent of 3- to 4-year-olds and 98 percent of 5- to 17-year-olds are enrolled in school.¹³ These figures are comparable to NSHAPC information about homeless children.

Homeless children live in households whose receipt of government benefits is quite similar to that of nonhomeless children in poor U.S. households. Seventy percent of children in homeless families receive food stamps, which does not differ from the 66 percent of poor U.S. children who do so. Nor does the proportion of both groups covered by Medicaid differ (73 percent of homeless and 69 percent of poor U.S. children). It is harder to tell whether differences exist between homeless and nonhomeless poor children in their family's receipt of cash assistance because the data are not reported in the same categories. Fifty-one percent of homeless children live in families receiving AFDC, and 12 percent live in families receiving SSI. The Bureau of the Census (1992, table E) reports that 55 percent of nonhomeless poor children live in households receiving cash assistance, which could be AFDC, SSI, General Assistance, or other means-tested cash benefits.

Combining homeless children with their homeless parent and with single homeless clients, table 2.3 shows how all homeless service users compare to the U.S. population in poverty on some basic demographic characteristics.¹⁴ All homeless service users include considerably more males than the overall American poverty population (65 versus 43 percent). They are less likely to be white (47 versus 67 percent) or Hispanic (11 versus 24 percent), and more likely to be black (42 versus 27 percent). They are also less likely to be children (24 versus 40 percent), less likely to be ages 55 and older (7 versus 15 percent), and more likely to be in their middle years (42 versus 19 percent are between the ages of 35 and 54).

The last issue of importance with respect to children of homeless clients is the question of where children are when they are not with their homeless parent. The answer is heavily dependent on the homeless clients' sex (figure 2.7). When the homeless client is male, his children who do not live with him are most likely to be with their (nonhomeless) mother (81 percent of male homeless clients' minor children). But only 23 percent of female homeless clients' minor children who do not live with their mother live with their father. The woman's own parents or other relatives are most likely to be caring for her children if they are not with her (46 percent of children of female homeless parents), and about one-fifth (19 percent of homeless women's minor children) are in foster care or group homes.

Service Needs, Stresses, and Vulnerabilities

Service Needs as Seen by Clients

Clients were asked to name the three things they needed most "right now," and also to identify the single most important thing keeping them in a homeless condition. Help finding a job was the most frequently cited need (42 percent), followed by help finding affordable housing (38 percent), and assistance with paying rent, mortgage, or utilities in relation to securing permanent housing (30 percent). Other needs cited by more than 10 percent of clients were assistance with transportation (19 percent), clothing (18 percent), food (17 percent), job training and medical care (13 percent each), and a GED and dental care (11 percent each).

Insufficient income was cited most frequently as "the single most important thing" keeping the client homeless, at 30 percent of all homeless clients (figure 2.8). An additional 24 percent cited lack of a job or employment. Lack of suitable housing was mentioned by 11 percent of homeless clients, and addiction to alcohol and/or drugs by 9 percent. No other categories except "other" (14 percent) were reported by many clients.

Food Consumption and Hunger

NSHAPC clients were asked a basic question about the sufficiency of the food they eat. Twenty-eight percent of homeless clients report that they sometimes or often do not get enough to eat. Only 39 percent get enough of the kinds of food they want to eat. This contrasts with 60 percent of poor U.S. households and 80 percent of all U.S. households who say they get enough of what they want to eat.¹⁵ Other food problems experienced by homeless clients include usually eating one meal a day or less (20 percent); being hungry in the past 30 days but not eating because one could not afford enough food (39 percent); and going a whole day without eating anything at all in the last 30 days (40 percent) (figure 2.9).

Data for some of these problems from poor U.S. households indicate that homeless clients have much higher levels of food problems than poor people generally. Thirty-nine percent of homeless clients versus 5 percent of poor households reported that in the last 30 days they were hungry but didn't eat because they couldn't afford to buy food, and 40 versus 3 percent said they didn't eat for one whole day or more because they couldn't afford to buy food.

An index of food problems based on clients' reported hunger and difficulties obtaining adequate food was also calculated. This five-level index reports the percentage of homeless clients reporting none, one, two, three, or four food problems. On this index, 42 percent of homeless clients report no food problems. By contrast, 37 percent have two or more food problems.

Victimization while Homeless

In addition to the stresses of finding enough food to eat, being homeless removes the safety of a permanent residence and leaves one's person and possessions vulnerable to attack. Robbery and theft are common threats experienced by two in five homeless clients (figure 2.10). Thirty-eight percent of homeless clients report having money or things stolen directly from them while they were present (robbery), and a similar proportion (41 percent) report having money or things stolen from their bags, locker, or other location while they were gone (theft). In addition, 22 percent report being physically assaulted or beaten up at least once while homeless, and 7 percent report being sexually assaulted or raped.

Physical Health Status and Insurance

Survey clients were given a list of 17 medical conditions and asked if they had any of them. These include conditions classified as

• acute infectious conditions (chest infection/cold/ cough/bronchitis, pneumonia, tuberculosis, STDs other than AIDS);
• acute noninfectious conditions (skin diseases, lice/ scabies); or
• chronic health conditions (diabetes, anemia, high blood pressure, heart disease/stroke, liver problems, arthritis/rheumatism, cancer, problems walking/lost limb/other handicap, HIV/AIDS), as well as "other" conditions.

Reported rates of these conditions may be underestimates because they rely on client self-reports. These self-reports may be low due to lack of knowledge or diagnosis of medical conditions, or reluctance to admit to having some of them. It is also possible that if the interview had inquired about other conditions, reported rates would be higher. Twenty-six percent of clients report one or more acute infectious conditions, 8 percent report one or more acute noninfectious conditions, and 45 percent report one or more chronic health conditions. Three of the four most commonly reported medical conditions are chronic health conditions: arthritis, rheumatism, or joint problems (24 percent); high blood pressure (15 percent); and problems walking, a lost limb, or other handicap (14 percent). Chest infection, cold, cough, or bronchitis (acute infectious conditions) are also among the most highly reported, at 22 percent.

Twenty-four percent of homeless clients report that they needed medical attention in the past year but were not able to get it. Forty-six percent could not get access to a dentist when one was needed. This lack of access may be due in part to their general lack of insurance coverage. Fifty-five percent report that they have no medical insurance of any kind; the comparable figure for all American adults is 16 percent. By contrast, 30 percent say they are covered by Medicaid, 7 percent by medical care through the Department of Veterans Affairs, 4 percent by private insurance, and 10 percent by insurance of other types. A few clients mentioned more than one type of insurance.

Among homeless clients in family households, 10 percent report that their children needed to see a doctor or nurse in the past year but were not able to do so. Homeless children are much less likely than homeless clients to be without insurance coverage. Only 20 percent of homeless family households have no insurance for their children. Seventy-three percent¹⁶ report Medicaid coverage for their children, while 6 percent have private insurance coverage and 6 percent have some other type of medical insurance coverage for their children (some clients in family households mentioned more than one type of insurance).

Alcohol, Drug, and Mental Health Problems

Clients were asked about experiences considered to be indicators of alcohol, drug, and mental health (ADM) problems, and about treatment experiences related to these problems. Responses to questions were combined to categorize clients as having or not having particular problems during the past month, past year, and/or their lifetime (see chapter 1 for details on how alcohol, drug, and mental health problems were defined). When looking at the results, it is important to remember that NSHAPC information is not based on clinical diagnoses.

Past Month. Homeless clients report a variety of problems with mental or emotional conditions, alcohol use, or use of illegal drugs within the past 30 days (table 2.4). During this period 38 percent report problems with alcohol use; 26 percent report problems with drug use, and 39 percent report mental health problems.

Sixty-six percent of homeless clients report one or more of these problems during the past month (figure 2.11). Thirteen percent report only alcohol problems, 7 percent report only drug problems, 17 percent report only mental health problems, 22 percent report combinations of two problems, and 8 percent report all three problems during the past month.

Past Year. The longer time period of the past year (including the past month) captures a larger proportion of homeless clients who report one or more problems with mental or emotional conditions, alcohol use, or use of illegal drugs. During this period, 46 percent report problems with alcohol use; 38 percent report problems with drug use, and 45 percent report mental health problems. Seventy-four percent of homeless clients report one or more of these problems during the past year (figure 2.11). Twelve percent report only alcohol problems, 7 percent report only drug problems, 15 percent report only mental health problems, 27 percent report combinations of two problems, and 14 percent report all three problems during the past year.

Lifetime. As the longest time period being considered, it is not surprising that lifetime histories reveal the highest level of problems. During their lifetime, 62 percent of homeless clients report problems with alcohol use; 58 percent report problems with drug use, and 57 percent report mental health problems. Eighty-six percent of homeless clients report one or more of these problems during their lifetime (figure 2.11). Nine percent report only alcohol problems, 6 percent report only drug problems, 10 percent report only mental health problems, 32 percent report combinations of two problems, and 30 percent report all three problems during their lifetime.

Overall Patterns. Looking over the three time periods examined, it is clear that as the time period lengthens more clients report problems in each area. Second, as the time period lengthens the proportion of clients who report a single problem decreases and the proportion who report combinations of two or three problems increases. Third, in each succeeding time period the difference decreases in the proportion reporting problems with alcohol compared to the proportion reporting drug problems, until for lifetime problems the difference is no longer statistically significant. Fourth, the proportion reporting combinations involving mental health problems plus alcohol and/or drug problems increases from 23 percent during the past month to 31 percent during the past year up to 47 percent over clients' lifetimes,¹⁷ with the most dramatic increase occurring in the proportion reporting all three.

History of Incarceration

About half (49 percent) of homeless clients have spent five or more days in a city or county jail in their lifetime. Some of these jail experiences may have been a direct result of their homelessness (i.e., the charges might be for behaviors that are difficult to avoid if one is homeless, such as loitering). Eighteen percent of clients have been in a state or federal prison, and 16 percent were held in juvenile detention at least once before reaching their 18th birthday. Altogether, 54 percent have some experience of incarceration.

Adverse Childhood Experiences

Ever since a Minneapolis study (Piliavin, Sosin, and Westerfelt 1993) identified childhood out-of-home placement in foster care as a common experience of homeless people, interest has been focused on these early separations from family and the ways that lack of family support after age 18 (when one has to leave foster care) might increase a young person's risk of homelessness. Homeless clients in the present study reveal that 27 percent were placed in foster care, a group home, or other institutional setting before their 18th birthday (figure 2.12). Many experienced multiple placements, as 12 percent were in foster care, 10 percent had been in a group home, and 16 percent had been in residential institutions.

Twenty-nine percent of homeless clients also report abuse or neglect in childhood from someone in their household (12 percent neglect, 22 percent physical abuse, and 13 percent sexual abuse). Thirty-three percent ran away from home and 22 percent were forced to leave home for at least 24 hours before they reached age 18. In addition, 21 percent report that their first period of homelessness predated their 18th birthday (this homelessness might have been with their family or on their own).

Where Homeless Clients Were Living

Even within the week documented by this study, homeless clients did not stay in one place. On the day of their interview and the seven days preceding it, many clients slept or rested in a number of different places. These could include places not meant for human habitation; emergency or transitional shelters; or temporary arrangements such as a house, apartment, or room in which someone is allowed to stay on a temporary basis.

Thirty-two percent of homeless clients slept or rested in places not meant for human habitation (designated "streets" in figure 2.13), including transportation depots, commercial spaces, cars or other vehicles, abandoned buildings, outdoor locations, and other venues of similar type. Thus just under one-third of homeless clients would have been found during a typical week's time using such places for sleep.

Homeless shelters are the most common type of location where homeless clients may be found. More than twice as many homeless clients (73 percent) slept in one or more of a variety of shelters during the eight-day period being examined as slept in places not meant for human habitation during the same period. Some, of course, slept in both types of venue. Shelters take many forms, including emergency shelters, transitional housing programs, and vouchers for emergency housing (designated "shelters" in figure 2.13).

Source: Urban Institute analysis of weighted 1996 NSHAPC client data.
^*Client used option at least once during the eight days including the day of the interview and the previous seven days, including being sampled at the site.
"Shelters"=emergency and transitional shelters and voucher programs; "Streets" =any place not meant for habitation; "Temporary Housing"= own or other person's house, apartment, or room, including hotel/motel room that client paid for, but without the possibility of sleeping there for the next month without being asked to leave.

In addition, figure 2.13 shows that 54 percent of homeless clients slept or rested in one or more temporary accommodations,18 including a friend's or relative's place, their own place, a hotel or motel room they paid for themselves, or a permanent housing program for formerly homeless people (designated "temporary accommodations" in figure 2.13).

To understand how extremely transient homeless clients are, it is important to examine the overlap in these categories. Six percent of homeless clients had slept or rested in all three venues during the eight-day period. At the other extreme, 7 percent had stayed only on the streets, 34 percent had stayed only in shelters, 6 percent had stayed only in temporary housing. The overlap is greatest for shelters and temporary accommodations, with 34 percent of homeless clients using both during the eight-day period. By contrast, the overlap between streets and shelters is quite low, with only 11 percent of homeless clients staying in both.

Characteristics of Clients Using Different Housing and Other Services

People staying in the different venues displayed in figure 2.13 are quite different from each other on a number of dimensions of importance to planners, service providers, and others offering assistance to homeless clients. To understand these differences, table 2.5 presents some basic descriptive information for clients who slept on the streets on the day of the interview and/or the previous seven days, and clients who during the same time period stayed in a shelter, used a soup kitchen, or used other types of homeless assistance programs. These groups are not mutually exclusive; many clients fall into more than one of them, and some clients could be included in all four.

Men are a similar proportion of clients who slept in the streets and those who used a soup kitchen during the last eight days (86 and 85 percent, respectively) and are more likely to be found in those venues than in shelters or other programs (where they comprise 65 and 67 percent of users). Few significant racial/ethnic differences exist between the four venues. Shelter stayers are more likely to be in a first homeless episode lasting six months or less (22 percent) than is true for any other group.

Table 2.5 also reports the prevalence of several subgroups within the four service use patterns. Relatively few survey clients in any venue are youth ages 24 and younger. Clients in homeless families are a larger proportion of shelter users (16 percent) and users of other programs (14 percent) than they are of street stayers (3 percent) or users of soup kitchens (6 percent). Persons reporting HIV/AIDS are a very small proportion (2 to 5 percent) in every venue.

More than half of clients in each venue have not done any work for pay in the last 30 days. Street stayers and users of other programs are less likely than shelter stayers to have done any paid work, and users of other programs are also less likely than soup kitchen users to have worked for pay during the past month.

The presence of any alcohol, drug, and/or mental health problems is highest among street stayers and those who used other programs (75 and 74 percent, respectively). Rates are lower but still considerable among homeless clients using shelters and soup kitchens (63 and 66 percent, respectively). Problems with drug use are lowest among shelter stayers (23 percent), and do not differ among clients using the other venues (31 to 35 percent). Both street stayers (45 percent) and other program users (43 percent) are more likely to report alcohol use problems than are shelter stayers (35 percent). The only significant difference by venue for mental health problems is that a higher proportion of users of other programs (46 percent) report such problems compared to those who stayed in a shelter (38 percent) or those who used a soup kitchen (37 percent).

These findings with respect to shelter and other program venues probably reflect the differing influences of program goals and rules. Many transitional housing programs are designed explicitly for clients with these problems, but many emergency shelters have rules that explicitly or effectively exclude clients with these problems, leaving the streets as their only alternative. Users of both types of programs are combined in the shelter user category. Further, some "other programs," especially outreach programs, drop-in centers, and permanent housing programs are explicitly designed to reach and serve clients with alcohol, drug, and/or mental health problems.¹⁹

Income, Employment, and Other Income Sources

Income

The average income of homeless clients during the month before being interviewed was $367 (table 2.6).²⁰ Further, 13 percent received no cash income at all during the past month. Clients in families averaged $475, but this amount had to support the parent and two children (on average). Homeless families thus were living on 46 percent of the federal poverty level of $1,023 for a family of three. Single homeless clients averaged $348 during the month before the interview, which was 51 percent of the federal poverty level of $680 a month for a single person. A comparison of these figures with the 1995 median monthly household income of $2,840 for all American households shows just how impoverished homeless clients really are.

Paid Employment

Almost half (44 percent) of homeless clients did some paid work during the 30 days before being interviewed, and 21 percent received money from family or friends. Of those who report working in the last 30 days, 20 percent did so in a job lasting or expected to last at least three months, 25 percent worked at a temporary or day labor job, and 2 percent earned money by peddling or selling personal belongings. Three percent name more than one source of earned income. Of those receiving money from family members or friends, 9 percent receive it from parents, 2 percent from a spouse, 5 percent from other relatives, 12 percent from friends, including boyfriends and girlfriends, and 1 percent from child support. Eight percent receive income from more than one type of friend or family member. Eight percent report obtaining money through panhandling. Considerably fewer (29 percent) homeless clients in families did any paid work. On the other hand, homeless clients in families were more likely than other homeless clients to receive money from family members or friends (32 versus 19 percent), including spouses, other relatives, friends including boyfriends and girlfriends, and child support. However, this help was clearly not enough to supply an adequate income.

Receipt of Government Benefits

Homeless clients receive income from a variety of sources in addition to earning income through a job. Means-tested government sources of income include AFDC (52 percent of homeless families), GA (9 percent), and SSI (11 percent). Food stamps was the government benefit helping the largest proportion of homeless clients (37 percent), followed by Medicaid (30 percent).

Receipt of means-tested benefits differs considerably by family status (figure 2.14). As would be expected, 52 percent of homeless families receive AFDC. In addition, homeless clients living in families are much more likely than single homeless clients to receive food stamps (61 versus 25 percent), and Medicaid (71 versus 31 percent). Although the two groups do not differ in their receipt of General Assistance or SSI, the difference attributable to the three programs linked through AFDC (AFDC, food stamps, and Medicaid) is enough to raise the level of homeless families receiving any means-tested benefit to 79 percent, compared with half of that (39 percent) among single homeless clients.

History of Homelessness

For almost half (49 percent) of homeless clients, their current spell of homelessness is their first (table 2.7). Of the remaining homeless clients, 17 percent are in their second spell and 34 percent have had at least three homeless spells including the current one. Twenty-eight percent of current episodes have lasted three months or less, another 11 percent have lasted between four and six months, 15 percent between seven and twelve months, 16 percent between thirteen and twenty-four months, and 30 percent have lasted two years or more.

Patterns of homelessness differ between homeless clients in families and single clients, primarily among those who are in their first homeless episode (figure 2.15). Both groups are equally likely to be in a first episode (50 and 49 percent). However, homeless families are more likely than single homeless clients to be in a short first episode (34 versus 15 percent), and less likely to be in a long first episode (16 versus 34 percent).

Homeless clients give many different reasons why they had to leave their last residence, but only a few reasons are identified by at least 5 percent of clients. These include not being able to pay the rent (15 percent), losing a job or having a job end (14 percent), doing drugs (7 percent), the landlord making one leave (6 percent), and not getting along with the people there (5 percent). Reasons for leaving one's last residence differ greatly by family status. More homeless clients in families than single homeless clients left because they could not pay the rent (22 versus 14 percent), because there was violence in the household (13 versus 2 percent), or because the landlord made them leave (12 versus 5 percent). Conversely, fewer homeless clients in families than single homeless clients say they left because they lost their job (2 versus 16 percent).

People Who Change Locations after Becoming Homeless

Fifty-six percent of homeless clients were interviewed in the same city, town, or rural community where they became homeless this time; that is, they have not changed communities since becoming homeless (table 2.7). An additional 22 percent say they have stayed for at least two days in two different communities since becoming homeless (including the one where they were interviewed), 8 percent have stayed in three different communities, and 14 percent have stayed in four or more different communities since becoming homeless. Among movers, 61 percent moved from one community to another within the same state, with 74 percent of homeless families and 59 percent of single homeless clients who moved staying within the same state.

Family status makes a big difference in the likelihood of moving from one community to another while homeless. Homeless clients in families were much more likely to have remained in the same community than were single homeless clients (71 versus 54 percent). In addition, they were much less likely to have stayed in five or more communities while homeless (1 versus 12 percent).

Why They Move. Respondents who report having left the community where this episode of homelessness began explained why they left that place, and why they came to the city or town where they were interviewed for NSHAPC. Many reasons were given for leaving their original town. The four most common were that there were no jobs in that place (18 percent), there was no affordable housing in that place (13 percent), they were evicted or asked to leave their housing (14 percent), and there were no services in that place (5 percent). Clients in families were less likely than single homeless clients to give lack of jobs as a reason for leaving (10 versus 19 percent), and more likely to give as reasons the lack of affordable housing (18 versus 12 percent), eviction/being asked to leave (20 versus 13 percent), and perceived lack of services (13 versus 4 percent).

Homeless clients who moved also gave many reasons for coming to the city or town where they participated in NSHAPC. Only four reasons were given by 10 percent or more of movers: they had friends and/or relatives here (25 percent), shelters/missions were here (21 percent), they were looking for work or heard there were jobs here (16 percent), and there were good services and programs here (19 percent). Only the two service-related reasons differentiate families from single homeless clients. Forty-one percent of families who moved mentioned the availability of shelters or missions in their new location, compared with 18 percent of single homeless who moved, while 27 percent of families who moved mentioned good services or programs as a reason to come to their current location, compared with 17 percent of single homeless clients who moved.

Where They Move from and Where They Move to. The basic pattern of moves between community types is for people to move to a place that is larger than the one they came from.²¹ Among people who changed the type of community in which they were living after becoming homeless for their current episode, 28 percent started in a large central city, 14 percent in a medium-sized central city, 31 and 10 percent, respectively, in the urban fringes of large and medium-sized central cities, 10 percent in large or small towns, 5 percent in rural areas, and 1 percent in another country (table 2.8).

Comparisons of Clients from Central Cities, Suburbs, and Rural Areas

Among all homeless clients, 71 percent were interviewed in central cities, 21 percent in suburban areas, and 9 percent in rural areas (table 2.9). This distribution does not vary by whether or not the client is in a family household, nor would it change if one included the children as well as their parents in the analysis. Homeless clients are thus much more likely to live in central cities than the U.S. poor population, 41 percent of whom live in central cities, 35 percent in the suburbs and urban fringe areas, and 23 percent in rural areas outside of metropolitan statistical areas (figure 2.16).

Central city and rural clients do not differ in their sex distribution (71 and 77 percent male), but suburban clients are significantly less likely to be male (55 percent). Suburban clients are also more likely to be white non-Hispanics (54 percent) than clients from either central cities (37 percent) or rural areas (42 percent).

In general, rural homeless clients have experienced fewer and shorter episodes of homelessness during their lifetimes. Fifty-five percent of rural clients have been homeless for three months or less, compared with 22 to 27 percent of central city and suburban homeless clients. In addition, only 27 percent have been homeless for more than a year, compared with 48 percent of central city and 49 percent of suburban clients.

A larger proportion of central city clients (21 percent) report staying in places not meant for habitation than is true for suburban (12 percent) or rural clients (4 percent). Clients from central cities are more likely than those from suburban/urban fringe and rural areas to have used a soup kitchen (68 percent, 50 percent, and 45 percent, respectively) and a drop-in center in their lifetime (30 percent, 18 percent, and 14 percent, respectively). The lack of availability of these programs outside of central cities most likely affects these results.

Central city homeless clients are considerably poorer than other homeless clients. Clients' median income is $250 in central cities, $395 in suburban areas, and $475 in rural areas. Fifteen percent of central city clients report no income over the last 30 days compared with only 6 to 7 percent of other clients.

Rural homeless clients report less access to medical care. Forty-seven percent of rural clients say they needed to see a doctor or nurse in the last year but were not able to do so, compared with 22 percent of homeless clients in both central cities and suburban/urban fringe areas.

Similar proportions (64 to 68 percent) of central city, suburban, and rural clients have a current mental health and/or alcohol and/or drug problem. However, clients from central cities are more likely (25 percent) than those in suburban areas (17 percent) to have current problems with both mental health and alcohol and/or drug use. And clients in both central cities and suburban areas are more likely than rural clients (11 percent) to have such problems.

Thirty-three percent of suburban homeless clients, 24 percent of those from central cities, and 12 percent of those from rural areas report being physically or sexually abused before the age of 18. Incarceration follows the opposite pattern, with the suburbs lowest and rural areas highest. Sixty-four percent of clients from rural areas have spent time in juvenile detention, jail, or state or federal prison, compared with 55 percent of those from central cities and 44 percent from suburban areas.

Comparisons with 1987 Urban Institute Findings

In 1987 the Urban Institute conducted a national survey of homeless clients using shelters and soup kitchens in large U.S. cities (those with 100,000 or more population in 1994). As there is considerable interest in examining whether, and in what ways, clients who are homeless may have changed during the nine years between the Urban Institute and NSHAPC surveys, relevant comparisons are presented here. For this analysis NSHAPC data have been restricted to be comparable to the 1987 data. This means that the 1996 statistics in this comparison are based only on homeless NSHAPC clients found in central cities who were sampled from shelters, voucher distribution programs, and soup kitchens. Table 2.10 presents comparable figures for 1987 and 1996 using these parameters to analyze the 1996 NSHAPC data.

As can be seen from the data in table 2.10, in 1996 homeless shelter and soup kitchen users located in central cities are less likely to be white (39 versus 46 percent) and more likely to be black (46 versus 41 percent) than those from 1987. They are more likely to have completed high school—39 versus 32 percent, and to have some education beyond high school—27 versus 20 percent. They are more likely to have married at some time in their lives (51 versus 45 percent), but have the same likelihood of being in a homeless family (10 percent in each year).

Although the mean length of clients' current homeless episode is shorter in 1996 than it was in 1987 (39 versus 32 months), there is no significant difference in the median length (12 versus 10 months).²² There is some increase in the proportion with short spells (30 versus 21 percent with spells of three months or less). However, there is no change in the proportion with long spells (31 percent of both 1987 Urban Institute and comparable 1996 NSHAPC clients reported spells of two or more years' duration).

In 1996, a larger proportion of homeless households received means-tested benefits (AFDC, SSI, and food stamps) than was true in 1987. In 1987, 33 percent of homeless clients in families reported receiving AFDC, which increased to 58 percent of comparable NSHAPC clients in 1996. Twice as large a proportion of all homeless central city users of shelters and soup kitchens received food stamps in 1996 compared with 1987 (37 percent versus 18 percent). For SSI the figures tripled (13 versus 4 percent). These differences are probably attributable to significantly greater efforts by homeless service providers to help clients obtain benefits, and to greater outreach efforts on the part of the government benefit programs themselves. Major SSI rule changes that effectively reduced the eligibility of many clients at high risk for homelessness (by eliminating eligibility for clients with a primary diagnosis of alcohol and/or drug abuse) did not take effect until after NSHAPC data were collected.

Probably as a consequence of increased access to public benefits, mean monthly per person income is higher in 1996 ($267) than it was in 1987 ($189)²³ among comparable clients. These estimates assume that the 10 percent of clients who have children with them have on average two children each, and that the remaining 90 percent of clients are by themselves.

The food situation has improved somewhat for homeless clients. In 1987, 38 percent of clients in the Urban Institute study (homeless users of central city shelters and soup kitchens) said they sometimes or often did not get enough to eat. Among comparable NSHAPC clients, 28 percent say the same in 1996. The greater likelihood of receiving food stamps may be related to this improved food situation. In addition, 31 percent say they get enough of the kinds of food they want to eat, compared with 19 percent in 1987.

The proportion of central city homeless clients who have experienced inpatient treatment for alcohol or drug abuse or for mental health problems has not changed at all between 1987 and 1996. The statistics in this analysis are based on all central city homeless clients, rather than being restricted to those who have any alcohol, drug, or mental health problem, because the 1987 Urban Institute data do not include the information needed to assess service use only among those with a problem.

---

Introduction

NSHAPC was designed to include interviews with all users of homeless assistance programs, including those who are not homeless. Information about nonhomeless clients helps in understanding who else is using these programs, and what experiences they may have with homelessness. In looking at results presented in this chapter that compare these groups, readers should be aware of some important limitations of NSHAPC data that limit simple inferences of causality. Formerly homeless clients and other service users in the NSHAPC sample are a random and representative sample of formerly and never homeless users of NSHAPC homeless assistance programs, but are not a representative sample of all formerly and never homeless clients in the United States.²

As will be seen, many similarities exist between currently and formerly homeless clients of NSHAPC homeless assistance programs (see chapter 1 for how these groups are defined). These similarities suggest two things, at least. First, some proportion of clients classified as formerly homeless by NSHAPC are essentially the same as currently homeless clients with a history of episodic homelessness, and the survey caught some in a homeless phase and others in a housed phase. Second, however, it is clear that some proportion of formerly homeless clients have been helped to leave homelessness through the auspices of public benefits and/or permanent housing programs. Because eligibility for these programs usually requires some level of disability (especially for programs that are federally funded), the inclusion of permanent housing program residents raises the level of reported health and other problems of NSHAPC's formerly homeless group.

Other users of homeless assistance programs are also of interest to service providers and policymakers. Although housed, their poverty is sufficient to bring them to homeless assistance programs for help, usually with food (75 percent were found in food programs, including soup kitchens, food pantries, and mobile food programs).

Basic Demographic Characteristics

Basic demographic characteristics for currently and formerly homeless clients and other service users are presented in table 3.1.

Sex

Men and women figure very differently in the three subpopulations of clients (figure 3.1). Men comprise 68 percent of currently homeless clients compared with 54 percent of formerly homeless clients. Among other service users 39 percent are men and 61 percent are women. Only the sex distribution of formerly homeless clients resembles that of the U.S. adult population, which was 48 percent male and 52 percent female in 1996 (Bureau of the Census 1997, table 14).

Race/Ethnicity

There are no significant racial differences between currently and formerly homeless clients. Equivalent proportions are white non-Hispanic (41 and 46 percent), black non-Hispanic (40 and 41 percent), Hispanic (11 and 9 percent), Native American (8 and 2 percent) and other races (1 and 2 percent). Among other service users, 54 percent are white non-Hispanic, 32 percent are black non-Hispanic, 11 percent are Hispanic, and 1 percent each are Native American and other races. However, all groups are significantly less likely to be white non-Hispanics than the U.S. population as a whole (76 percent), and more likely to be black non-Hispanics (40, 41, and 32 percent versus 11 percent) (Bureau of the Census 1997a, table 23).

Age

Young clients (ages 17 to 24) make up a greater share of currently than formerly homeless clients (12 versus 4 percent) (figure 3.2). The age group of 25- to 34-year-olds is also more common among currently than formerly homeless clients (25 versus 17 percent). The opposite is true as age increases; 45- to 54-year-olds as well as those ages 55 and older are more numerous among formerly than currently homeless clients. Sixty-one percent of other service users are ages 45 and older, with 29 percent being ages 65 and older.

Education

No differences exist in the educational attainment of currently and formerly homeless clients. Thirty-four percent of each group are high school graduates or have a G.E.D., with 28 and 24 percent, respectively, having some education beyond high school. About half of other service users (49 percent) have not completed high school, 32 percent have high school diplomas, and 19 percent have some education beyond high school. All three groups are similar in their proportion of high school graduates compared with the U.S. adult population as a whole (at 34 percent). But all U.S. adults are significantly more likely than any group of NSHAPC clients to have some education beyond high school (48 percent) (Bureau of the Census 1997a, table 245).

Marital and Household Status

Similar proportions of currently and formerly homeless clients have never married (48 and 45 percent), divorced (24 and 25 percent), and separated (15 and 14 percent). In addition, equal proportions are married (9 percent in each group) and widowed (3 and 6 percent). Among other service users, 28 percent have never married, 22 percent are currently married, 25 percent are widowed (reflecting their high proportion of older women), 10 percent are separated, and 15 percent are divorced.

Information about whether or not a client lives in a family was combined with information about the client's age and sex to describe as best as possible the client's household status. The results are reported in table 3.1.

Among clients ages 17 to 24, about one-fourth of currently homeless clients (3 of 12 percent) and one-sixth of formerly homeless clients (1 of 6 percent) are women living in families. Currently homeless clients are more likely than formerly homeless clients to be single men ages 25 and older (61 versus 49 percent). Ten percent of other service users are ages 17 to 24, among whom 4 in 10 are women living in families. Sixteen percent of other service users are clients ages 25 and older living in families. Single clients ages 25 and older comprise 74 percent of other service users; 33 percent are men and 41 percent are women.

Veteran Status

The proportion of currently and formerly homeless clients who are veterans is similar (23 and 22 percent, respectively). Among other service users the proportion is 14 percent. The Department of Veterans Affairs estimates that in 1995, 13 percent of all American adults were veterans.

Service Needs, Stresses, and Vulnerabilities

Service Needs as Seen by Clients

Each client was asked to select from an extensive list of various needs "the three things you need the most help with now." Responses are reported in table 3.2.

Currently compared with formerly homeless clients are more likely to report needing help finding a job (42 versus 30 percent), help finding affordable housing (38 versus 21 percent), and assistance with rent, mortgage, or utility payments for securing permanent housing (30 versus 15 percent). Formerly homeless clients are more likely to say they need help obtaining food (32 percent, compared with 17 percent for currently homeless clients). Among other service users, the highest service needs are for help obtaining food (33 percent), and other needs (31 percent).

"Other" needs were mentioned by almost one-third of other service users, and by one-fourth of currently and formerly homeless clients. Small proportions mentioned needing help with getting clothing, transportation, legal aid, medical or dental care for themselves or their children, drug/alcohol/mental health treatment, enrolling children in school, and domestic violence, in addition to assistance with job training and managing money. Also many clients mentioned an "other" need that was personal, such as "a good man/woman," "spiritual growth," and "peace of mind." All of these responses are included in the "other" response category.

Food Consumption and Hunger

Clients reported whether they (1) get enough of the kinds of foods they want; (2) get enough but not always the kinds of foods they want; (3) sometimes do not get enough food to eat; and (4) often do not get enough food to eat. Currently and formerly homeless clients report a similar likelihood (28 and 25 percent) of sometimes or often not getting enough to eat (table 3.3). This likelihood is much higher than the likelihood of having similar difficulties among all U.S. households (4 percent) and even among all poor households in the country (12 percent).³

Currently and formerly homeless clients are equally likely to report two or more problems getting enough food (38 and 31 percent). Thirteen percent of other service users report this level of problems getting enough food. Individual problems over the past 30 days included in this index are eating once a day or less, being hungry but not having money for food, and going a whole day without eating because of not having money for food. Similar proportions of currently and formerly homeless clients report eating once a day or less (20 and 17 percent). Currently homeless clients are more likely than formerly homeless clients to report the other two measures (39 versus 26 percent for "hungry but not enough money for food," and 40 versus 33 percent for "whole day without eating"). Both are much higher than parallel figures for all U.S. households (2 and 1 percent, respectively) and poor U.S. households (5 and 3 percent, respectively).

Other service users report some level of difficulty getting enough food. Sixteen percent report sometimes or often not getting enough to eat, and 13 percent report two or more food-related problems. Specific problems reported for the past 30 days include 10 percent who report eating once a day or less, 11 percent who were hungry but did not have enough money for food (compared to 5 percent of poor U.S. households), and 17 percent who went a whole day without eating because they could not afford food (compared to 3 percent of all U.S. households). Thus other service users also have more problems getting adequate food than do poor U.S. households.

Physical Health Status and Insurance

Survey clients were asked about 17 medical conditions, classified subsequently as acute infectious conditions (chest infection/cold/cough/bronchitis, pneumonia, tuberculosis, STDs other than AIDS); acute noninfectious conditions (skin diseases, lice/scabies); chronic health conditions (diabetes, anemia, high blood pressure, heart disease/stroke, liver problems, arthritis/ rheumatism, cancer, problems walking/other handicap, HIV/ AIDS); whether they used drugs intravenously; and other medical problems.

Currently and formerly homeless clients report the same levels of acute infectious or acute noninfectious conditions. However, formerly homeless clients report higher rates of chronic health conditions than do currently homeless clients (62 versus 46 percent). Two-thirds of other service users report chronic conditions, 29 percent report acute infectious conditions, and 3 percent report the acute noninfectious conditions about which the survey asked (table 3.4).

The most common conditions among all groups are arthritis, rheumatism, and joint problems, although at quite different levels. Currently homeless clients are less likely to report these chronic conditions than are formerly homeless clients (24 versus 37 percent). Currently homeless clients are also less likely than formerly homeless clients to report high blood pressure (15 versus 22 percent), and problems walking or other physical handicap (14 versus 22 percent). Among other service users the rate for arthritis and similar problems is 42 percent, high blood presssure is 33 percent, problems walking/other handicap is 30 percent, and upper respiratory problems is 27 percent.

Similar proportions of currently and formerly homeless clients (24 and 26 percent) needed but were unable to see a doctor in the year prior to the survey (figure 3.3). This figure is much lower among other service users (12 percent). Far fewer of the children living with clients were affected by this lack of access to care, with similar proportions (8 and 9 percent) of parents in each subgroup reporting any problems with access for their children.

Coverage by health insurance varies considerably by homeless status (figure 3.4). Clients could report more than one type of insurance, so figures do not sum to 100 percent. Over half (55 percent) of currently homeless clients have no health insurance, compared with 32 percent of formerly homeless clients. More formerly homeless clients than currently homeless clients receive Medicaid (53 versus 30 percent). Other service users report health insurance coverage by Medicaid (43 percent), private insurance (13 percent), other insurance (28 percent), and no insurance (31 percent). The "other" insurance reported by other service users is probably Medicare, given the age of many in this group (28 percent are ages 65 and older).

Similar proportions (73 and 74 percent) of currently and formerly homeless clients living in families report that their children are covered by Medicaid. (Clients could report more than one type of insurance coverage.) Only 6 to 7 percent of the children living with currently and formerly homeless clients are covered by private insurance. Twenty percent of currently homeless households with children have no health insurance for their children, compared with 14 percent of formerly homeless households with children. These high levels of health insurance coverage for children living with NSHAPC parents help to explain why parents report that relatively few of their children experienced any problems getting needed health care.

Alcohol, Drug, and Mental Health Problems

Measures of alcohol, drug, and mental health (ADM) problems are reported for past month, past year (including past month), and lifetime (including past year). These measures are composites of client reports of behaviors, experiences, and treatment, as defined in chapter 1.

In general, for alcohol and drug problems in all three time frames, currently homeless clients report higher rates than formerly homeless clients. Currently and formerly homeless clients report similar rates of mental health problems in all time periods (table 3.5). Looking at combinations of alcohol, drug, and mental health problems indicates that for relatively recent time periods, currently homeless clients have higher rates of one or more problems during the past month (66 percent) and past year (74 percent) than formerly homeless clients (59 and 66 percent for past month and past year, respectively). However, when the probability of having at least one of these problems over clients' lifetimes is examined, currently and formerly homeless clients are equally likely (86 and 87 percent) to have had at least one of the problems (figure 3.5).

Looking at combinations of different problems experienced by the same person, as the time period lengthens from past month to lifetime, more currently and formerly homeless clients report at least one ADM problem (table 3.5). Also, for the longer time periods (past year and lifetime), formerly homeless clients are more likely than currently homeless clients to report only mental health problems (21 versus 15 percent for past year; 17 versus 10 percent for lifetime). Alcohol use problems are more prevalent than drug use problems in each time period for both currently and formerly homeless clients. Alcohol use and mental health problems show the same levels in all time periods for currently homeless clients, but formerly homeless clients report more problems with mental health than with alcohol use for past month (41 versus 29 percent) and past year (46 versus 33 percent). Currently, compared with formerly, homeless clients report higher rates of alcohol use problems during the past month (38 versus 29 percent) and past year (46 versus 33 percent), but are similar over their lifetimes (62 versus 56 percent).

Thirty-one percent of other service users report at least one ADM problem during the past month, 36 percent do so for the past year, and 56 percent do so for lifetime problems. Mental health and alcohol use problems appear in this group at the same levels for past month (16 percent each) and past year (17 and 21 percent), but lifetime alcohol use problems are higher (36 versus 28 percent). Drug use problems are less commonly reported (4 percent for past month, 7 percent for past year, and 18 percent for lifetime).

Incarceration

Currently and formerly homeless clients are equally likely (49 and 43 percent) to have spent five or more days in a city or county jail, which may be related to their condition of homelessness as well as to other behavior (table 3.6). More currently than formerly homeless clients report having spent time in a state or federal prison (18 versus 9 percent) or in juvenile detention before they reached the age of 18 (16 versus 9 percent). Taking all of their incarceration experiences together, 54 percent of currently homeless clients have spent some time incarcerated, compared with 45 percent of formerly homeless clients. Only 14 percent of other service users have ever been incarcerated.

Adverse Childhood Experiences

NSHAPC results indicate that mental health and alcohol and/or drug problems may have roots in the childhood of many clients, as does homelessness itself. Between one-fifth and one-fourth of both currently and formerly homeless clients report that before the age of 18 they (1) started drinking and using drugs; (2) experienced physical abuse, sexual abuse, or both from someone in their household; (3) spent time in juvenile detention; (4) lived in foster care or other out-of-home placement; (5) ran away from or were forced to leave home; or (6) became homeless for the first time (table 3.7).

In general, currently homeless clients are more likely than formerly homeless clients to report initiation of drinking and using drugs at younger ages. Analysis of data from the National Household Survey on Drug Abuse indicates that people who start drinking and using drugs before the age of 15 are at significantly higher risk for problems with substance use as adults (Dennis and McGeary 1998). NSHAPC data indicate that 25 percent of currently homeless clients began drinking before age 15, and that 19 percent first started using drugs at this young age. Both of these rates are higher than those for formerly homeless clients (19 percent for drinking and 13 percent for using drugs) (figure 3.6).

Twenty-five percent of currently homeless clients and a similar proportion of formerly homeless clients (22 percent) report being physically and/or sexually abused by a household member before reaching their 18th birthday (figure 3.6). Rates that combine neglect with abuse experiences show roughly the same pattern, but for this combination currently homeless clients do report significantly higher rates than formerly homeless clients (29 versus 23 percent). Only 8 percent of other service users report any abuse or neglect experience before age 18 (table 3.7).

Finally, 16 percent of currently homeless clients had spent some time in juvenile detention before they reached the age of 18. This proportion is higher than that for formerly homeless clients (at 9 percent). Other service users report a rate of 4 percent for time spent in juvenile corrections institutions. Rates of experiences away from one's childhood family reported by currently and formerly homeless clients are high enough and often similar enough to suggest, as other research has done, that these may be significant risk factors for adult homelessness (e.g., Piliavin, Sosin, and Westerfelt). These questions were not asked of other service users.

More currently than formerly homeless clients (27 and 19 percent, respectively) were placed in foster care, a group home, or other out-of-home placement (figure 3.7). Similar proportions of both groups had been forced to leave home for more than 24 hours (22 and 19 percent, respectively). The two groups differ, however, in the proportion who had run away from home for more than 24 hours (33 percent of currently and 27 percent of formerly homeless clients) and had a homeless experience before the age of 18 (21 and 12 percent). Their answers indicate that many do not consider their runaway experience to be homelessness, since many more in both groups report the former than the latter experience.

Income Levels, Sources, and Employment

Currently homeless clients report a lower level of income during the past 30 days, on average, than formerly homeless clients ($367 versus $470, table 3.8).⁴ Further, 13 percent of currently homeless clients report no cash income at all during the past month, compared with 5 percent of formerly homeless clients. Both are significantly below the federal poverty level of $680 a month for a single person. Median incomes are very low, at $300 for currently homeless and $462 for formerly homeless clients. Other service users report a mean income over the past 30 days of $575, and a median of $514.

Currently homeless clients are more than twice as likely as formerly homeless clients to have incomes below $100 during the past 30 days (30 versus 14 percent). The proportion of clients with incomes below $300 is 49 percent for currently homeless clients, 30 percent for formerly homeless clients, and 21 percent for other service users (figure 3.8).

Currently, compared with formerly, homeless clients are more likely to have worked for pay during the last 30 days (44 versus 34 percent) (table 3.8). However, although they are less likely to work at all, a higher proportion of the work activity of formerly homeless clients is in jobs lasting or expected to last three months or more, whereas more than half of the work done by currently homeless clients is in temporary or day labor jobs. Forty-five percent (20/44 = 45%) of currently homeless clients who work report jobs lasting or expected to last three months or more, compared with 62 percent (21/34 = 62%) of formerly homeless clients with earned income.

Food stamps are the most common source of means-tested government benefits for currently and formerly homeless clients (37 and 48 percent, respectively) (figure 3.9). Currently homeless clients are less likely than formerly homeless clients to receive SSI (11 versus 29 percent). Formerly homeless clients are more likely than currently homeless clients to receive General Assistance (16 versus 9 percent). Formerly homeless clients are also more likely than currently homeless clients to receive SSDI (16 versus 8 percent). Other service users are most likely to get food stamps (37 percent), Social Security (33 percent), and SSI (26 percent).

History of Homelessness

Currently and formerly homeless clients were asked about their experiences with homelessness, including how many times they had been homeless, the length of their current homeless episode (if homeless now), and the length of the most recent completed episode of homelessness (for formerly homeless clients and currently homeless clients who have been homeless more than once). Results are shown in table 3.9.

Currently homeless clients are more likely than those who are formerly homeless to have had only one homeless episode (49 versus 40 percent). Similar proportions of currently and formerly homeless clients (22 and 21 percent) have been homeless four or more times (figure 3.10).

Twenty-eight percent of currently homeless clients have been homeless during their present spell for three months or less, while 30 percent have been homeless for two years or more. There are no differences between formerly homeless clients and those among the currently homeless who have had more than one spell regarding the length of their most recent completed spell. The most frequently mentioned spell length was between one and three months (33 and 30 percent, respectively). The next most frequently mentioned category was episodes of 7 to 12 months, indicated by 21 and 20 percent. Relatively few clients reported completing episodes of two years or more (15 and 12 percent).

Use of Homeless Assistance Programs

Clients' use of various homeless assistance programs, including food and other programs in addition to shelters, is reported in table 3.10. Program use within the week before being interviewed for NSHAPC shows significant variation by homeless status. Soup kitchens are the most commonly used program among currently and formerly homeless clients (31 and 35 percent, respectively, report using them at least once during the week before the interview), although the difference between soup kitchen, emergency shelter, and transitional housing program use is not significant for currently homeless clients. Emergency and transitional shelters were each used by about a quarter of currently homeless clients (26 and 28 percent) but, by definition, were not used by formerly homeless clients and other service users. Other service users also report the most use of soup kitchens (19 percent).

Lifetime Service Use

The proportion of clients reporting having ever used various programs or services at some time in their lives is obviously much higher than usage in the week before being interviewed (table 3.10). Equal proportions (65 percent) of currently and formerly homeless clients have used an emergency shelter at some time in their lives. More formerly than currently homeless clients (19 versus 10 percent) have used permanent housing programs (this statistic probably owes a lot to the fact that using the program is part of what gets a person classified as formerly homeless), while the situation is reversed for transitional housing programs (40 versus 23 percent). The two groups report similar access to programs offering vouchers for emergency accommodation (15 and 18 percent).

Formerly compared with currently homeless clients are more likely to have used a soup kitchen in their lifetime (74 versus 62 percent). Formerly homeless clients are more likely than currently homeless clients to have used food pantries (59 versus 40 percent). Mobile food programs have served equal proportions (21 and 18 percent) of currently and formerly homeless clients. Similar proportions of currently and formerly homeless clients have used drop-in centers (26 and 27 percent) and outreach programs (17 and 14 percent). Other service users frequent food pantries (50 percent) and soup kitchens (32 percent), with lower percentages using mobile food programs (17 percent), drop-in centers (11 percent), and outreach programs (9 percent). Use of food programs of all varieties clearly reflects great need related to food among every subgroup of clients included in NSHAPC.

Urban/Rural Location

The urban/rural location of currently and formerly homeless clients and other service users is dramatically different (table 3.11). Currently homeless clients are more likely than formerly homeless clients to be found in central cities (71 versus 64 percent), but the proportion for both is higher than the proportion of the U.S. population in poverty found in these locations. Currently compared with formerly homeless clients are less likely to be found in rural areas (9 versus 19 percent). Other service users are equally likely to be found in central cities and in rural areas (39 and 40 percent).

---

Introduction

Over the past decade the number and variety of programs serving homeless clients have grown tremendously, but no national description of this service network has been available. NSHAPC provides the first national description of this service network, covering central cities, suburbs and urban fringe communities, and rural areas. This chapter offers an overview of programs and services available in the United States that have a significant focus on serving homeless clients, spanning a wide variety of program types (exhibit 4.1). Readers interested in a more detailed examination of programs and services are referred to chapters 14 through 17 of Homelessness: Programs and the People They Serve—Technical Report.

Service Locations and Program Types

Estimates from NSHAPC data indicate there are approximately 21,000 service locations in the United States, operating about 40,000 homeless assistance programs. A little over half of the service locations (about 11,000) offer only one homeless assistance program, and a little less than half offer two or more programs. For purposes of this study, a service location is the building or physical space at which one or more programs are offered, and a homeless assistance program is a set of services offered to the same group of clients at a single location (see chapter 1 for a detailed discussion of definitions).

Three features of NSHAPC's definition of a program are important to keep in mind throughout this chapter: (1) nonhomeless clients often use some of these programs; (2) NSHAPC programs are not the only sources of assistance to homeless clients; and (3) many programs similar to NSHAPC programs may exist in a community but have not been included because they do not target their services toward homeless clients (food pantries and health programs are examples).2

Figure 4.1 shows the estimated number of NSHAPC programs by four general program types of housing, food, health, and other programs. Note that financial/ housing programs (e.g., Emergency Food and Shelter Program, welfare, public housing programs) were not an original NSHAPC category, but were mentioned frequently enough under "other programs" to warrant presentation as a separate category.

Food pantries are the most numerous type of program serving homeless clients, with an estimated 9,000 programs nationwide. Emergency shelters are next, with almost 5,700 programs, followed closely by transitional housing programs (4,400), soup kitchens and other distributors of prepared meals (3,500), outreach programs (3,300), and voucher distribution programs (3,100). As a group, homeless assistance programs with a health focus are least numerous.

Source: Weighted NSHAPC data representing programs operating during "an average week in February 1996."
Note: Financial (usually welfare) and housing assistance services were mentioned often enough under "Other" to warrant a category of their own, and migrant housing was combined with "Other" because there were so few programs.

Service Level and Program Size

Of considerable interest to many people are estimates of the number of clients being served by homeless assistance programs. NSHAPC results offer an important overview of service utilization in the United States, but they must be interpreted correctly. Since people may use more than one type of service during an average day, the estimates of service levels made by NSHAPC programs necessarily contain an unknown and unknowable amount of duplication. Their answers cannot be added up to determine the total number of clients who use services on an average day. For that reason they are referred to as "program contacts," not as "clients served." Further, many of the clients using the programs included in NSHAPC are not homeless, so care must be taken to interpret the following figures appropriately, as simple estimates of program use by clients who may need a wide variety of different services and may get them at a variety of programs.

A final caveat is that these figures do not represent all of each type of service available to homeless clients, for two reasons. First, shelters serve meals to their residents, and may also offer some types of health and other services. Therefore, NSHAPC food and health programs do not represent all of the food and/or health services available to homeless clients within a community. Second, some homeless clients get services from mainstream housing, health, and social service programs that were not included in NSHAPC program universe because they do not meet the study's definition of a homeless assistance program.

Source:Weighted NSHAPC data representing programs operating during "an average week in February."
Note: These are estimates from NSHAPC program representatives of how many contacts their program expected on an average day in February 1996. They contain duplication and cannot be added together to get the total number of people served on an average day.

Total Number of Expected Service Contacts

For each type of NSHAPC program, figure 4.2 shows how many contacts the programs expected to have on an average day in February 1996. Food pantries as a group clearly expected to have the most program contacts (over 1 million) on an average day, followed by about 520,000 contacts at soup kitchens. Programs offering financial and/or housing assistance, outreach programs, and emergency shelters each expected to have between 240,000 and 250,000 program contacts a day. In contrast, the estimate of program contacts for all four types of health programs with a focus on serving homeless clients, taken together, is only about 140,000, and this estimate is accurate only if each person uses one and only one type of health service on an average day. These figures probably are high estimates for average daily service use, since February is a peak month for many homeless assistance services, and program representatives tend to recall their peak periods rather than their average days.

Source: Weighted NSHAPC data representing programs operating during "an average week in February 1996."
Note: These are program staff estimates of how much program contacts their own program expected on an average day in February 1996. They contain duplication and cannot be added together to get the total number of people served on an average day. Housing programs include emergency, transitional, permanent housing, and voucher programs; food programs include pantries, soup kitchens, and mobile food programs; health programs include general health, mental health, alcohol/drug, and HIV/AIDS programs; other programs include outreach, drop-in centers, financial/housing assistance, and other.

Variability in Program Size

Homeless assistance programs vary greatly in size, where size is defined as the number of service contacts expected on an average day in February 1996. Figure 4.3 shows how programs of different types are distributed by size. The figure makes clear that food programs are most likely to be quite large (26 percent expected 101 to 299 service contacts daily and 11 percent expected more than 300 service contacts daily), and only about 1 in 11 expected as few as 1 to 10 service contacts in a day. Shelter and housing programs are likely to be small (28 percent expected 1 to 10 and another 31 percent expected 11 to 25 service contacts a day), with only 2 percent expecting more than 300 service contacts daily. Health and other homeless assistance programs are the most evenly distributed across a range of sizes, with about 40 percent of each expecting 25 or fewer and between 44 and 45 percent of each expecting between 26 and 100 service contacts daily.

The biggest programs, though few in number, account for very large proportions of the clients being served on an average day (figure 4.4). This is true regardless of which type of program one examines, but is most true for shelter/housing programs. The 80 percent of shelter/housing programs serving 50 or fewer clients daily serve only 32 percent of all the clients who use these programs on an average day (in February 1996). On the other hand, the 8 percent of shelter/housing programs serving more than 100 clients daily serve 51 percent of the clients using shelter/housing programs. Indeed, the 2 percent of these programs serving more than 300 clients daily serve 28 percent of all shelter/housing users on an average day.

The story is the same with food and other homeless assistance programs. Only 11 percent of food programs and 5 percent of other programs serve more than 300 clients daily, but these programs accommodate, respectively, 55 percent of everyone getting food from food programs and 55 percent of everyone getting help from other programs on an average day. Service delivery in health programs for homeless clients is less skewed toward the very large programs (over 300 daily) and away from the very small programs (25 or fewer daily), but even here the 42 percent of programs that are very small serve only 7 percent of those who use health programs on an average day, while the very large programs serve 30 percent of health program users.

Source: Weighted NSHAPC data representing programs operating during "an average week in February 1996."
Note: These are program staff estimates of how much program contacts their own program expected on an average day in February 1996. They contain duplication and cannot be added together to get the total number of people served on an average day. Housing programs include emergency, transitional, permanent housing, and voucher programs; food programs include pantries, soup kitchens, and mobile food programs; health programs include general health, mental health, alcohol/drug, and HIV/AIDS programs; other programs include outreach, drop-in centers, financial/housing assistance, and other.

Operating Agencies

Nonprofit agencies offer the vast majority (85 percent) of homeless assistance programs (figure 4.5). Secular nonprofits offer 51 percent of all programs, while religious nonprofits offer 34 percent. Government agencies operate 14 percent of all programs, and for-profit firms account for a mere 1 percent.

Different types of programs are operated by quite different types of agencies (figure 4.6). Secular nonprofit agencies dominate in the housing category, offering 60 percent of all programs, while religious nonprofits dominate in the food category, offering 55 percent of these programs. Health programs are about evenly split between government and nongovernment agencies, with secular nonprofits dominating among the nongovernment agencies that offer health programs for homeless clients.

Government agencies are least likely to operate food programs (5 percent of those programs) and most likely to offer health programs (51 percent of health programs). Secular nonprofits are the most prominent type of agency among other programs offering homeless assistance, which include outreach programs, drop-in centers, housing and financial assistance programs, and other programs. For-profit organizations play almost no role in operating homeless assistance programs (1 percent in housing, health, and other programs, and nothing in food programs).

Funding Sources

Given the different types of agencies operating homeless assistance programs, it should not be surprising that considerable differences also exist across program types in the extent to which they rely on private funding or funding from government sources (federal, state, or local) to support their operations. Figure 4.7 shows these differences as the proportion of program budgets that come from government funding; private funding makes up the balance of each program's budget.

Food programs are the most likely to rely solely on private funding, and to report that they use no government funds (true for 51 percent of food programs). They are also the least likely to be fully supported by government funds (6 percent). Conversely, health programs are the least likely to rely solely on private funding sources (only 12 percent operate entirely without government funding) and the most likely to be fully supported from government sources (55 percent).

In between are housing and other types of homeless assistance programs. One-fourth of all housing programs for homeless clients are fully supported by government funds, and about equal proportions rely entirely on private funds (23 percent) or receive up to half of their budget from government sources (22 percent). The final 30 percent of housing programs receive from half to almost all of their support from government funds. Other homeless assistance programs are split quite evenly among the one-third that rely entirely on private funding, the 34 percent that are completely supported by government, and the one-third whose level of government support falls somewhere in between.

The value of in-kind contributions from private sources are not included in these figures. For many programs, these contributions can be of considerable value and include food, rent-free buildings, equipment, and volunteer time to perform critical program functions.

Population Focus

Some homeless assistance programs are open to anyone who wants to use them, while other programs are designed specifically to serve only certain types of people. Population focus may be defined in several different ways. One common way is by household type, including men by themselves, women by themselves, households with children, and youth by themselves. Other common ways are by the special population groups, such as veterans, youth, victims of domestic violence; or condition or service need, such as alcohol, drug, or mental health problems, or HIV/AIDS, that a program is specifically designed to help. NSHAPC program staff were asked to report each household type that their program serves. In another series of questions, program representatives described any specializations or particular focuses that their program might have. This section reports how programs describe the types of clients they serve on these different dimensions. Since many programs report serving more than one household type, and also report having a focus on more than one special population, condition, or service need, the statistics for each type of program usually sum to more than 100 percent.

Household Type

Figure 4.8 reports the types of households that programs serve. A very noticeable pattern is that food programs are the most inclusive of all household types, with 84 percent or more saying they serve men and women by themselves and both female-headed and other families with children. Housing programs are the most specialized. For instance, 43 percent serve two-parent families with children, 61 percent serve men by themselves, and 68 to 69 percent serve women by themselves and female-headed households with children. Of course, many serve more than one of these household types. These patterns may be due to program policy (i.e., the program will not take particular types of clients) or simply to the fact that few or no clients of a particular type come to the program for service.

Nine out of ten health programs serve men and women by themselves, but only six in ten serve any families with children. Shelter and other housing programs are the least likely to expect unaccompanied youth to use their programs, but all of the other program types also do not expect to serve many of this group.

Special Population or Special Need

Respondents were asked whether their program had a particular focus on one or more special populations. These might include a population group, such as victims of domestic violence, runaway youth, and veterans; or special conditions or special needs, such as mental health problems (without alcohol and/or drug use), alcohol and/or drug use (without mental health problems), both alcohol/drug and mental health problems, and HIV/AIDS.

Half of those naming any special focus named only one focus, 17 percent named two focuses, and 33 percent named three or more focuses. The most commonly named special focus was "other" (25 percent). About half (48 percent) of program representatives did not report any special focus; food programs were the most likely and health programs were the least likely to say they had no special focus.

Figure 4.9 identifies special population groups reported as a program's focus, and figure 4.10 does the same for program focus with respect to condition or service need. Overall, figure 4.9 shows that one-third or fewer programs named a special population group as a focus. The group most frequently named by all types of programs is victims of domestic violence, followed by veterans (18 and 14 percent, respectively, of all programs, shown in the first bar in each cluster). There are some important variations by program type, with housing programs being the most likely to name victims of domestic violence as a special population focus (28 percent do so), and health programs being most likely to identify veterans as a special population focus (32 percent do so). No more than 10 percent of programs identify a special focus on youth, regardless of program type.

Figure 4.10 shows the proportion of programs responding that they have a focus on serving clients with particular special needs. Overall, people with alcohol, drug, or mental health problems, alone or in combination, are a special focus for 17 to 19 percent of programs. Health programs are by far the most likely to report these focuses. Almost half of health programs say they have a special focus on clients with mental health problems only, and almost half report a special focus on clients with mental health problems plus accompanying alcohol and/or drug use. More than one-third of health programs have a special focus on persons who use alcohol or drugs, and slightly less than one-third of health programs have a special focus on clients with HIV/AIDS.

Occupancy Levels of Shelter/Housing Programs

Shelter/housing programs generally report occupancy levels in the 70 percent range (table 4.1). Programs serving homeless families report 73 percent occupancy for most of the year. The slightly higher winter rate of 77 percent for these programs is statistically but probably not substantively significant. Programs serving single homeless men and/or women report occupancy levels of 67 to 69 percent in the spring, summer, and fall, and significantly higher occupancy in the winter (76 percent). The three reasons given most frequently for being less than full are decline in need, seasonal changes, and that the people seeking shelter do not meet program criteria/restrictions. Given that many programs appear in both columns of table 4.1 since they serve both families and single men and/or women, it is not surprising that their reasons for less-than-full occupancy levels are similar.

Location of Homeless Assistance Programs

About half (49 percent) of all homeless assistance programs are found in central cities. Rural communities offer the next largest share of programs (32 percent), and suburban/urban fringe communities offer the smallest share of programs (19 percent) (figure 4.11). However, because central city programs serve more clients, on average, a larger share of program contacts happen in central cities (57 percent) than in suburban and rural areas (20 and 23 percent of all program contacts, respectively).³

Source: Weighted NSHAPC data representing programs operating during an "average week in February 1996." Includes all programs, of every type.

Central city, suburban, and rural locations also vary considerably in terms of which programs report the most contacts. Food programs report the most contacts in both central cities and suburban areas (57 and 65 percent of program contacts in those areas, respectively), while in rural areas food program contacts comprise only 31 percent of program contacts (figure 4.12). The share of program contacts that occur in shelter/housing programs are similar for central city and suburban areas (23 and 21 percent, respectively), but are only 11 percent of contacts in rural areas. In contrast, contacts with "other" programs predominate in rural areas (50 percent of contacts), but make up only 16 percent of contacts in central cities and 11 percent of contacts in suburban areas.

Availability of Services within NSHAPC's 76 Sampling Areas

So far, this chapter has explored homeless assistance programs and services at the national level and through the broad geographical designations of central cities, suburban areas, and rural areas. NSHAPC also has the capacity to examine variations in programs and program contacts within the survey's 76 primary sampling areas (called "sampling areas" hereafter).⁴ Program contact information can be segmented to reveal the proportion of services within a sampling area that are shelter/housing program contacts, food program contacts, health program contacts, and other program contacts. Further, program contact information can be used to calculate a rate of program contacts per 10,000 clients in poverty, which is a good measure for comparing the level of service availability across sampling areas of very different sizes (e.g., a city with more than a million people and a rural area of a few thousand people).⁵

Source: Weighted NSHAPC data representing programs operating during "an average week in February 1996." Includes all programs, of every type.

Source: Weighted NSHAPC data representing programs operating during "an average week in February 1996." Housing programs include emergency, transitional, permanent housing, and voucher programs; food programs include pantries, soup kitchens, and mobile food programs; health programs include general health, mental health, alcohol/drug, and HIV/AIDS programs; other programs include outreach, drop-in centers, finacial/housing assistance and other.

Distribution of Program Contacts within Sampling Areas

All other things being equal, one might expect the sampling areas with the largest population to provide the most homeless assistance program contacts. To examine this expectation, figure 4.13a arrays each of this study's 76 primary sampling areas from left to right according to the size of its population (largest on the left, smallest on the right).⁶ Each bar shows the total estimated number of program contacts on an average day in February 1996,⁷ combining program contacts of all types.

From this figure one can see that the expected relationship of more program contacts in the larger sampling areas is generally true but there are exceptions. The average estimated number of program contacts per sampling area is about 17,600 on an average day in February 1996. But the estimated numbers of program contacts range from a high of about 186,000 to a low of nothing (for two sampling areas that had no programs of any kind) (table 4.2). And variation exists even at the highest end. For example, providers in the largest sampling area estimate only about two-thirds the number of program contacts (about 123,000) as do providers in the next largest sampling area (about 186,000).

To examine how much population size accounts for the differences observed in figure 4.13a, the estimated number of program contacts per 10,000 people is employed. Figure 4.13b shows this rate for each of the 76 sampling areas, arrayed in the same order as figure 4.13a. The average estimated rate of program contacts per 10,000 population in a sampling area is 122. The use of a common denominator (10,000 people) reduces the differences among sampling areas quite a bit. Now one can see that some of the smaller sampling areas in the middle and toward the right of the figure appear to provide more units of homeless assistance services per capita than do some of the largest sampling areas.

Yet another way to look at these data is to ask whether the variability in service levels can be accounted for by the size of a sampling area's population in poverty, and not just by the total number of people in the sampling area. There is some reason to expect that services should be related to poverty, and the number of poor people in a sampling area is the best measure of need that is available for all 76 sampling areas. Some sampling areas could have a lot of people but not very many poor people, while some smaller sampling areas might actually have more poor people than some larger areas. Therefore a second rate was constructed for each sampling area—its rate of program contacts per 10,000 poor people. Figure 4.13c shows the results.

The average estimated rate of program contacts per 10,000 poor people in a sampling area is 1,437. The rate of contacts per 10,000 poor people reduces even further the level of variability in service provision among the largest sampling areas at the left of the graph. The variability in the middle of the graph (medium- and small-sized metropolitan areas) appears to have increased in relation to that in figure 4.13b. The mostly rural areas to the right of figures 4.13b and 4.13c appear to have the greatest variability whichever rate is used.

From the three graphs in figure 4.13 one can draw the conclusion that the biggest sampling areas, which comprise the nation's biggest cities, do not always provide the most services on a per capita basis, even though they obviously provide very large numbers of services. One can also conclude that a great deal of intercommunity variability remains in the provision of homeless services, even after controlling for levels of population and poverty. This level of variability probably stems from important differences in philosophies, policies, resources, and experience among communities.

Distribution of Services within Sampling Areas by Program Type

The next issue to be examined is how the total estimated number of program contacts within each sampling area are distributed among the four major program types of shelter/housing, food, health, and other. The results, shown in figure 4.14, reveal great variation in the proportion of service contacts in sampling areas within shelter/housing, food, health, and other program types.

Figure 4.14a-d shows four panels, one each for shelter/housing, food, health, and other program contacts. The average proportion of program contacts reported by shelter/housing programs is 24 percent, by food programs is 49 percent, by health programs is 5 percent, and by other programs is 19 percent (table 4.2 gives means, highs, lows, and standard deviations). In comparing parts (a) through (d) of figure 4.14, one can see the predominance of food program contacts and the relative paucity of health program contacts. Food program contacts comprise at least 40 percent of all program contacts in most sampling areas (only 17 of the 74 areas with any services have less than 40 percent of their program contacts at food programs, and one-third have more than 60 percent of program contacts at food programs). In contrast, only five sampling areas have as much as 20 percent of program contacts occurring at health programs, and most have less than 10 percent in the health area.

The greatest variability occurs in smaller metropolitan areas and rural areas, which are the most likely to have either much more of a concentration in a particular type of service than is true nationally, or much less of a concentration. Some of these sampling areas have all or virtually all of their program contacts in housing programs, others have all or almost all their contacts in "other" programs (such as outreach, drop-in, or housing/financial assistance programs), and a few have a significant share in health programs.

Distribution within Sampling Areas of Contacts with Different Types of Shelter/Housing Programs

Shelter/housing program distributions in sampling areas reflect very different decisions about where to invest homeless housing resources. This analysis uses a rate of shelter/housing program contacts per 10,000 poor people. Figure 4.15 provides this information, first for all shelter/housing program types (figure 4.15a), and then separately for each type of shelter and housing program (emergency shelter—figure 4.15b; transitional housing—figure 4.15c; permanent housing for the formerly homeless—figure 4.15d; and vouchers for temporary shelter—figure 4.15e).

The estimated national rate of program contacts with all types of shelter and housing programs for homeless people is 195/10,000 poor people. In addition to the two sampling areas with no programs of any kind, one additional sampling area has no shelter/housing program contacts at all.

Emergency shelter contacts per 10,000 poor people in the study's primary sampling areas average 81/10,000. Six sampling areas offer 150 or more shelter/housing contacts per 10,000 poor people, while nine sampling areas offer 20 or fewer emergency shelter contacts per 10,000 poor people including four that do not offer any.

Variability is even greater among primary sampling areas for rates of transitional housing, permanent housing, and voucher distribution. Transitional housing contacts within sampling areas have a national average of 49/10,000 poor people, with eight sampling areas offering more than 100 transitional housing contacts per 10,000 poor people and 23 offering 20 or fewer, including 13 that offer none.

Permanent housing contacts within sampling areas have a national average of 40/10,000 poor people, with seven sampling areas offering more than 100 permanent housing contacts per 10,000 poor people and 42 offering 20 or fewer, including 20 that offer none.

Voucher distribution contacts within sampling areas have a national average of 26/10,000 poor people, with four sampling areas offering more than 100 voucher program contacts per 10,000 poor people and 50 offering 20 or fewer, including nine that offer none.

---

NSHAPC offers the first opportunity since 1987 to examine homeless assistance programs and their clients across the nation. This landmark survey provides nationally representative data about the providers of homeless assistance and the characteristics of currently homeless and other persons who use services, information that is vital to national discussions about homelessness.

The study reveals that the level and type of homeless assistance programs across America are as diverse as homeless people themselves. The homeless people these programs serve have very limited income and other resources, and a complex array of needs. And homeless people are present in rural areas as well as urban and suburban locations.

The study also allows valuable comparisons between 1987 and 1996 of the characteristics of homeless people using shelters and soup kitchens in central cities. It shows that the already high percentages of racial and ethnic minorities using these services in 1987 became even higher by 1996. Although educational levels, income, and receipt of a variety of means-tested government benefits were higher in 1996 than in 1987, extreme poverty remained a central fact of life for homeless clients, whose income was generally half or less of the federal poverty level.

Since 1996, there have been major changes to the national, state, and local social welfare systems that are reportedly having significant impacts on America's low-income people. The impacts on homeless persons, both positive and negative, need to be identified. Comparing the findings of the 1987 and 1996 studies would provide important guidance for needed changes to homeless programs and mainstream social welfare programs.

In a similar manner, communities around the nation will need to conduct their own studies to guide local homeless and mainstream policy decisions. The methodology used in this study provides procedures and questionnaires that are readily adaptable for local surveys. Such surveys would offer local policymakers the twin advantages of having their own data and being able to see how local programs and service users compare to those in the nation as a whole.

Finally, future studies might consider specific objectives—such as the effectiveness of homeless assistance programs or estimates of the number of homeless people—that this survey was not designed to address.

---

APPENDIX A.2

Primary Sampling Areas Sorted by Population Size, in Descending Order

• Los Angeles—Long Beach, CA
• New York, NY
• Chicago, IL
• Philadelphia, PA-NJ
• Boston, MA-NH
• Detroit, MI
• Washington, DC-MD-VA-WV
• Houston, TX
• Atlanta, GA
• Dallas, TX
• Nassau-Suffolk, NY
• Riverside-San Bernardino, CA
• Minneapolis-St. Paul, MN-WI
• San Diego, CA
• St. Louis, MO-IL
• Orange County, CA
• Pittsburgh, PA
• Baltimore, MD
• Phoenix-Mesa, AZ
• Cleveland-Lorain-Elyria, OH
• Oakland, CA
• Tampa-St. Petersburg-Clearwater, FL
• Seattle-Bellevue-Everett, WA
• Miami, FL
• Newark, NJ
• Denver, CO
• San Francisco, CA
• Kansas City, MO-KS
• Norfolk-Virginia Beach-Newport News, VA-NC
• Indianapolis, IN
• Sacramento, CA
• Bergen-Passaic, NJ
• Salt Lake City-Ogden, UT
• Oklahoma City, OK
• Birmingham, AL
• Springfield, MA
• Youngstown-Warren, OH
• Melbourne-Titusville-Palm Bay, FL
• Shreveport-Bossier City, LA
• York, PA
• Utica-Rome, NY
• Boise City, ID
• Savannah, GA
• Lincoln, NE
• Abbeville County, Greenwood County, Laurens County, McCormick County, Newberry County, Saluda County, SC
• Bremerton, WA
• Bedford County, Fulton County, Huntingdon County, Juniata County, Mifflin County, PA
• Christian County, Clay County, Effingham County, Fayette County, Montgomery County, Moultrie County, Shelby County, IL
• Bradford County, Columbia County, Dixie County, Hamilton County, Lafayette County, Madison County, Suwannee County, Taylor County, Union County, FL
• Hancock County, Hardin County, Putnam County, Wyandot County, OH
• Jackson, MI
• Redding, CA
• Bangor, ME
• Las Cruces, NM
• Kenosha, WI
• Chaves County, Lea County, NM
• Dover, DE
• Buena Vista County, Clay County, Dickinson County, Emmet County, O'Brien County, Osceola County, Palo Alto County, Pocahontas County, IA
• Burnett County, Clark County, Rusk County, Sawyer County, Taylor County, Washburn County, WI Haskell County, Latimer County, Le Flore County, Pittsburg County, OK
• Douglas County, OR
• Iredell County, NC
• Crisp County, Dooly County, Macon County, Marion County, Schley County, Sumter County, Taylor County, Webster County, GA
• Cibola County, McKinley County, NM
• Caldwell County, Daviess County, Grundy County, Harrison County, Linn County, Livingston County, Mercer County, Putnam County, Sullivan County, MO
• Bath County, Menifee County, Montgomery County, Morgan County, Rowan County, KY
• Hall County, Hamilton County, Merick County, NE
• Aransas County, Bee County, Live Oak County, Refugio County, TX
• Enid, OK
• Accomack County, Northampton County, VA
• Lassen County, Modoc County, CA
• Houston County, Humphreys County, Stewart County, TN
• Johnson County, Sheridan County, WY
• Chester town, Deep River town, Essex town, Lyme town, Westbrook town, CT
• Union Parish, LA
• Esmeralda County, Mineral County, NV