The Ecology of Homelessness

Abstract

This article proposes an ecological model of homelessness drawn from a broad review of the research and practice literature. The spectrum of biopsychosocial risks associated with pathways into homelessness is reviewed, followed by examination of social and individual consequences resulting from periods of homelessness. The temporal dimensions of homelessness, including “first-time,” “episodic,” and “chronic,” are defined and discussed along with the continuum of living arrangements experienced by homeless individuals and families. These constituent domains of the ecology of homelessness are presented in a conceptual map depicting the relationships and interactions among the parts of the model. The intention is to provide a coherent and cogent map to encourage comprehensive integrated efforts as individuals, agencies, and communities strive to prevent and resolve homelessness.

KEYWORDS:

• Homelessness
• ecological model
• biopsychosocial risks
• housing status
• temporal dimensions

Introduction

From an ecological perspective, homelessness can be understood as the result of interactions among risk factors ranging from individual conditions to socio-economic structures and environmental circumstances (Toro, Trickett, Wall, & Salem, 1991; Baron, 2004; National Coalition for the Homeless, 2007). Homelessness manifests itself on a temporal continuum as situational, episodic, or chronic. Over time, homeless individuals may experience changes in housing status that include being on the street, shared dwelling, emergency shelter, transitional housing, and permanent housing and hospitalization and incarceration in correctional facilities. Episodes of homelessness result in individual and social consequences, which are commonly detrimental to individual well-being and negatively affect social interactions within the community (Phelan, Link, Moore, & Stueve, 1997; Saelinger, 2006; National Coalition for the Homeless [NCH], 2004; National Law Center on Homelessness and Poverty [NLCHP], 1997.)

The intention of this article is to propose a broad conceptual model of homelessness that examines biopsychosocial risk factors associated with homelessness in relation to the constructs of temporal course, housing status, and individual and social outcomes. We employ an ecological perspective to situate and describe known biopsychosocial risk factors in a hierarchy of systems/domains. The goal is to transcend the classic debate that posited homelessness as the result of either individual or structural factors (Burt, Aron, Lee, & Valente, 2001; Fisher & Breakey, 1991; Calsyn & Roades, 2006; United States Conference of Mayors, 2007). Though framing the debate of the origins of homelessness within this artificial causal dichotomy may have served political and policy objectives, this reductionism does not advance an etiological understanding of homelessness reflective of the phenomenon's actual complexity nor does it foster robust, multi-systemic response options from communities, agencies, organizations, and practitioners. Moreover, homelessness cannot be understood or addressed by focusing solely on causal factors and ignoring its varying temporal dimensions, the spectrum of consequential individual and social outcomes, and the resulting limited housing options associated with homelessness. To view homelessness only from the perspective of why and how individuals and families became homeless is to see only half the picture. The conceptual model presented here is intended to articulate a gestalt of homelessness not only recognizing the constituent parts of the phenomenon but offering a map of the dynamic interactions of the elements of the model.

The proposed model is consistent with recognition of the complexity of homelessness. There are increasing attempts to analyze the transactional nature of factors contributing to homelessness (Burt et al., 2001; NCH, 2007; Coolen, 2006; Martijn & Sharpe, 2006; Toro et al., 1991). In our model, the four primary components are biopsychosocial risk factors, individual and social outcomes, the temporal dimension, and housing outcomes. Figure 1 illustrates this model of the ecology of homelessness. It is intended to depict the dynamic relationship between the domains and elements of the model.

BIOPSYCHOSOCIAL RISK FACTORS

Biopsychosocial risk factors encompass a range of factors including individual biology and development and circumstances such as poverty and its many facets to housing availability and stability (NCH, 2007). The concept of biopsychosocial implies an ecological perspective. It recognizes the interaction of multiple factors on different levels, including individual factors (e.g., personality, developmental experiences, health-mental health, race, and ethnicity) and social factors, such as resource availability, policies, culture, discrimination, and social situations. It moves away from dichotomies such as micro vs macro or individual versus structural to appreciate the continuous transactions between person and environment. The concept underscores the complexity of interactions on different systems levels and encourages analysis of homeless as resulting from individual and family risks or vulnerability within a social context.

FIGURE 1 Ecological model of homelessness.

Structural

Poverty

Poverty is overriding and intertwined in homelessness. People without financial resources are unable to meet basic needs such as housing and food, nor can they obtain other needed services. Poverty is a risk factor that makes people vulnerable for homelessness:

… poverty represents a vulnerability, a lower likelihood of being able to cope when the pressure gets too great. It thus resembles serious mental illness, physical handicap, chemical dependency, or any other vulnerability that reduces one's resilience and the resilience of one's family and friends. (Burt, 1992)

A number of factors contribute to poverty including unemployment, the declining value of the minimum wage, housing costs, and health care and other services (NCH, 2007; US Conference of Mayors, 2005).

Employment and the minimum wage

Lack of employment is often identified as a major cause of homelessness; however, many homeless persons report being employed or having occasional work (Economic Policy Institute, 2005; US Conference of Mayors, 2005). The difficulty is that many of these jobs do not provide adequate wages and benefits for self-sufficiency, a trend spanning a number of years. Mishel, Bernstein, & Schmitt (1999) recognized that the value of the minimum wage had not kept up with economic growth. The United States Interagency Council on Homelessness (USICH) found that the median monthly income for persons who were homeless was about 44% of the federal poverty level (1999). Though the value of the minimum wage has not kept up with inflation, there has also been a decline in manufacturing jobs, a corresponding increase in low-paying service employment, globalization, decline in union bargaining power, and increase in temporary work, all factors in wage decline (USICH, 1999).

Many of the temporary jobs held by homeless persons do not provide sufficient wages and benefits such as health insurance to ensure self-sufficiency. The Interagency Council on the Homeless (ICH, 1999) recognized that employment prospects are dim for those who lack appropriate skills or adequate schooling. The labor market has changed, as evidenced by “plant relocations and closures, persistent racial discrimination, changes in industry that have increased the demand for highly educated people, the decline in the real value of the minimum wage, and the globalization of the economy” (ICH, p. 27). Employment instability has been identified in several studies as a risk factor for homelessness (Wagner & Perrine, 1994). Women and minorities seem to experience fewer employment opportunities (Anti-Discrimination Center of Metro New York, 2005; American Civil Liberties Union [ACLU], 2004; Butler, 1997). Furthermore, the duration of homelessness may decrease the prospects of employment. It is not surprising that homelessness itself may further diminish one's chances of employment, as prolonged idleness may cause loss in work habits, responsibility, and commitment to employment.

Loss of public benefits

The decline in availability of public assistance and its declining value is a risk factor for homelessness. The welfare reform legislation in the late 1990s resulted in a number of individuals and families losing benefits. The Institute for Children and Poverty found that more than one-third of homeless families had benefits reduced (2001). The loss of health insurance also increases vulnerability (Families USA, 2001).

Welfare reform, including the shift from aid to families with dependent children (AFDC) to the block grant program temporary assistance to needy families (TANF) and the reduction in eligibility for SSI based on chronic substance abuse, has had an especially devastating impact. A study of TANF benefits, for example, found that assistance to a single mother of two children provided an income at approximately 29% of the federal poverty level (Nickelson, 2004). In essence, as homeless families experience reduced benefits, additional barriers to escaping homelessness result (Institute for Children and Poverty, 2001).

Housing costs and availability

The shortage of affordable, particularly rental, housing is a major risk factor for homelessness. Approximately 2.2 million low-rent units were lost between 1973 and 1993 owing to abandonment, conversion to condominiums, or becoming unaffordable because of competition and costs (Daskal, 1998). The Institute for Children and Poverty (2001) estimated a gap of more than 4 million units between affordable units and low-income renters. Quigley and Raphael (2001) indicate that individuals on the lowest end of the economic scale are extremely vulnerable to homelessness secondary to shifts in the housing market.

Nationally, urban change and policy initiatives in the United States and global changes have contributed to a decline in affordable housing (National Low Income Housing Coalition, 2005). The loss of single room occupancy (SRO) housing has been particularly devastating (NCH, 2007). Dolbeare (1996) estimates that more than one million units were lost in the 1970s and 80s. For example, in 1960, 640,000 people in New York lived in SROs and rooming houses, but by 1990, this number was reduced to 137,000 (Stegman, 1993). Most cities have witnessed private-sector hotels and rooming houses that provided cheap lodging being razed or converted to condominiums in the apparent gentrification of the inner city (Wright & Rubin, 1997). In smaller urban areas, it may be that the “new SROs” are the increasing number of suburban motels, formally tourist-focused but now offering low rates and catering to a transient population (Nooe, 2006).

As affordable housing units and especially SROs were lost, the competition for (and rent levels of) remaining units increased. Those with mental illness and addiction disorders were significantly impacted in terms of not being able to locate and afford housing. As the NCH (2007) points out, in the last two decades, competition for increasingly scarce low-income housing grew so intensely that those with disabilities such as addictive and mental disorders were more likely to lose out and find themselves on the streets.

Family housing instability

Families with children are the most rapidly growing group of the homeless population (US Conference of Majors, 2007). The lack of affordable housing, exacerbated by low incomes of families often with a history of housing instability, is a major factor in family homelessness (Burt, 2001; Shin et al., 1998). Housing instability involves a number of factors such as the lack of income, unavailability of supportive housing, and discrimination. It often manifests in doubling up with other families and living in shelters and cars (United States Department of Housing and Urban Development [USDHUD], 2009).

Homeless families are most frequently headed by single mothers (Rog & Buckner, 2007). A number of studies have found that female-headed households have greater risks for poverty (United States Department of Commerce, 1999) and subsequently have greater risks of homelessness (Caton et al., 1995; DiBlasio & Belcher, 1995). Similarly, women who have experienced violence may encounter discrimination from landlords who are reluctant to rent to them (ACLU, 2004).

Deinstitutionalization

According to a survey by the US Conference of Majors (2005), approximately 16% of the single adult homeless population has a severe and persistent mental illness. However, a number of studies suggest a higher incidence, perhaps one-third of the homeless suffering severe mental illness (ICH, 1994; Nooe, 2006; Task Force on Homelessness and Severe Mental Illness, 1992). The estimated rates are wide ranging depending on methodology, definitions, sample selection, and diagnostic criteria; for example, shelter users seem to have higher rates of mental illness than do non-sheltered homeless persons. The Annual Homeless Assessment Report to Congress (USDHUD, 2009) indicted that 25% of all sheltered persons have a disability condition. Though the report does not specify the specific type of disability, severe mental illness and chronic substance abuse are frequently reported.

The relationship of homelessness to mental illness was initially highlighted when Bassuk asked the question, “Is homelessness a mental health problem?” (Bassuk, Ruben, & Lauriat, 1984). Underlying this deceptively simple question are a number of issues about individual causes versus structural causes, including the role of deinstitutionalization. Discussion of the influence of mental illness on homelessness has continued in the literature (Bassuk et al., 1984; Baum & Burnes, 1993; North, Pollio, Smith, & Spitznagel, 1998; HCH Clinicians' Network, 2000).

The role of deinstitutionalization in homelessness is a complex one that is compounded by the lack of national data on the outcomes of persons discharged from mental institutions. While deinstitutionalization increased through the 1980s, federal policies reduced funding for social programs with transfer of fiscal responsibility to state and local governments. Concurrently, conservative public housing policies reduced the availability of low-income housing, increasing competition, and the mentally ill became at greater risk for homelessness (Jansson & Smith, 1996; NCH, 2008). The importance of housing availability is also underscored by the decline in persons discharged from institutions returning to live with families (Talbott, 1980).

In summary, the history of the deinstitutionalization movement is fairly clear; however, there continues to be debate about its causal relationship to homelessness. Torrey (1989) argued:

The homeless mentally ill are a product of the best intentions followed by the worst of operations. They are the result of deinstitutionalization, the policy that evolved in the 1950s and 60s to shift care of the seriously mentally ill from state mental hospitals to community facilities.

The NCH (Mental Illness and Homelessness, 1997) initially asserted that deinstitutionalization was not a major factor but more recently identified it as a contributing factor (2008).

The foregoing illustrates the different viewpoints. Most likely deinstitutionalization is a contributing factor but not the sole cause. Failure to provide support to those deinstitutionalized and the reduced availability and accessibility of resources for both treated and untreated persons are factors. For those persons with a mental illness, homelessness has a detrimental effect and, like any other crisis or trauma, may “catalyze and/or exacerbate mental illness, producing disorder where previously it did not exist” (Koegel & Burnam, 1992, p. 96).

Health care costs

The unavailability of health care is a critical factor for homelessness. Faced with the challenge to pay rent and secure food and other basic necessities, persons with serious illness or disability can find themselves in a crisis that ends in homelessness. Between 2005 and 2006, the number of people without health insurance coverage increased from 44.8 million (15.3%) to 47 million (15.8%; US Bureau of the Census, 2007). More than a third of persons living in poverty have no health insurance. In recognizing this risk factor, the National Health Care for the Homeless Council calls for the establishment of a national health plan that guarantees access to affordable, high-quality and comprehensive health care, which is essential in the fight to end homelessness (NHCHC, 2008). In essence, health care costs may increase the risk for homelessness while lack of access to health care may be heightened by homelessness.

Low wages

The study, Homelessness: Programs and the People They Serve (US Interagency Council on Homelessness, 1999) found that median monthly income for persons who were homeless was about 44% of the federal poverty level. The value of the minimum wage, as noted previously, has not kept up with inflation. Decline in the value of minimum wage with the parallel decreased availability, but increased competition for affordable housing has heightened the risk for homelessness.

Employment prospects are especially dim for those who lack appropriate skills or adequate schooling. The labor market has changed, as evidenced by “plant relocations and closures, persistent racial discrimination, changes in industry that have increased the demand for highly educated people, the decline in the real value of the minimum wage, and the globalization of the economy” (US Interagency Council on Homelessness, 1999). Women and minorities are especially vulnerable and seem to experience fewer employment opportunities and lower wages (Anti Discrimination Center of Metro New York, 2005; ACLU, 2004; Butler, 1997). Employment instability is a risk factor for homelessness (Wagner & Perrine, 1994) and may create a cyclical phenomenon. Thus, homelessness itself may decrease the prospects of employment (NCH, 2008).

Discrimination

The foregoing discussion recognizes that women and minorities often experience discrimination that manifests in lower wages or discriminatory practices and increases the risk of homelessness. People experiencing homelessness experience further stigmatization and discrimination in a range of contexts including access to health/mental health care, education, employment, and shelter (NCH, 2004; Phelan et al., 1997). Laws can operate in a manner that disadvantages homeless persons. The report “Illegal to Be Homeless: The Criminalization of Homelessness in the United States” describes violations to the basic human rights to homeless persons through passage of possibly unconstitutional laws, selective enforcement, and incorporation of discriminatory regulations (NCH, 2004). Ordinances prohibiting large group feedings, for example, have been found to be discriminatory against the homeless (ACLU, 2008). These discriminatory practices may involve criteria for Social Security, banning activities in public spaces, or more subtle denial of housing or employment without a fixed address or owing to past history (e.g., domestic violence, history of incarceration; ACLU, 2004; Anti-Discrimination Center of Metro New York, 2005; Human Rights and Equal Opportunity Commission, 2008; Tolman, Danziger, & Rosen, 2001).

The foregoing discussion has focused on risk factors that are structural or represent the social context within which the individual or family functions. Interacting with these factors are risk factors that can be viewed as individual.

Individual

Individual risk factors include personal characteristics or conditions such as age, marital status, race, mental illness, substance abuse, and educational level and personal experiences such as childhood trauma and foster care that increase vulnerability for homelessness. These factors tend to imply personal responsibility. These risk factors are integrated and interacting as reflected in Figure 1, rather than representing linear causation. The point for emphasis is that they often increase vulnerability for homelessness.

Age

Age can be considered a factor in homelessness both in children, as discussed earlier in family homelessness, and also in older persons, generally those older than 50 years of age. The number of older persons living in poverty and becoming homeless has increased, and the demand for affordable housing for elderly persons exceeds the supply (US Census Bureau, 2007). Elderly persons falling below the poverty level face greater challenges in maintaining housing and food, and unavailability of health care often creates a crisis situation that ends in homelessness. Once on the street, the elderly homeless may experience exacerbated health problems (Cohen, 1999) and being more vulnerable to crime and exploitation (NCH, 2007). It is not surprising that the premature mortality rate is three to four times greater for the elderly homeless (O'Connell, 2005).

Marital status

The question of the relationships between marital status and homelessness has been addressed in a number of studies (AHAR, 2007; Burt et al., 1999; Rossi, 1989). Forty-eight percent of homeless individuals have never married. The Annual Homeless Assessment Report to Congress (2007) noted that the majority of shelter users were unaccompanied adults. Among those who have been married (52%), 39% have ended in divorce or separation. Earlier discussion noted research evidence that single female-headed households were more vulnerable to homelessness, in that homeless families are most frequently headed by single mothers (Rog & Buckner, 2007). Jencks observed that “married couples hardly ever become homeless as long as they stick together” (1994). Marriage may provide social support, sharing of responsibilities, and pooling of resources. Though marital status may not be causal, it is likely a risk factor that affects vulnerability to homelessness and influences the duration of homelessness (Calsyn & Roades, 2006).

Social support

The research literature indicates homeless persons typically have smaller social networks than the non-homeless (Calsyn & Winter, 2002; Shinn, Knickman, & Weitzman, 1991). Bassuk et al. (1997) report that 220 homeless mothers in family shelters in Worcester, MA had significantly fewer members in their social networks than a comparison group of housed women. Further, the homeless women reported more conflict in their relationships than the housed women. This finding is echoed in an Anderson and Rayens (2004) study of 255 women, of whom 98 were homeless, 88 had never been homeless but experienced childhood trauma (physical and/or sexual abuse), and 73 had never been homeless and did not experience childhood trauma. When compared to the two groups of non-homeless women, the homeless women had significantly lower levels of social support and significantly higher levels of relationship conflict.

Letiecq, Anderson, & Koblinsky (1998) found homeless mothers in emergency shelters and in transitional housing had significantly less social support, as evidenced by less help from families, fewer people to count on, and lower levels of weekly contact with friends and relatives, than a control group of housed mothers. Kingree, Stephens, Braithwaite & Griffin (1999) found that after completion of a substance abuse treatment program, low levels of support from friends were associated with homelessness. Similarly, adolescents running away from or being kicked out by families are at risk for homelessness (Maclean, Embry, & Cauce, 1999). The availability of ongoing support for those exiting foster care, mental health, and correctional facilities is especially critical for avoiding or escaping homelessness.

Foster care

Several studies of homelessness have found that a history of foster care may be a childhood precipitant to becoming homeless (Roman & Wolfe, 1997). In a study of 220 homeless mothers, Bassuk et al. (1997) found the two most important childhood predictors of family homelessness were foster care placement and drug use by primary female caretakers. Though there is an over-representation of persons who have been in foster care among the homeless, the casual relationship is not conclusive. Most likely foster care and other childhood disruptions represent risk factors. Roman and Wolfe (1997) suggest that foster care may also be associated with homelessness at an earlier age and a longer duration of homelessness. Among children “aging out” of foster care, estimates suggest as many as 22% become homeless within a year (Pecora et. al., 2005). There is some evidence that remaining in foster care until 21, as compared to aging out at 18, reduces the risk.

Family conflict and violence

In Congressional testimony, Burt points out that “Very high proportions of homeless youth report family conflict as a reason for being homeless” (2007, p. 5).

Family conflict and family breakdown are frequently identified by homeless youth as contributing to their homelessness (Mallett, Rosenthal, & Keys, 2005). The pathways leading to youth homelessness often involve strained family relationships, conflict, communication problems, and parental substance abuse or mental health problems (Rosenheck, Bassuk, & Salomon, 2008). In a qualitative study of homeless youth, Miller, Donahue, Este, and Hofer (2004) found that in descriptions of families of origin, violence was common, whereas nurturance and support were lacking or absent.

Though family conflict is often cited as a contributing factor to youth homelessness, it is also a factor in homelessness among households with children and among singles. A survey of cities conducted by the US Conference of Mayors (2007) found that family disputes were reported as a cause of homelessness by 17% for both households and singles. Though not all family conflict involves domestic violence, studies have indicated that as many as 50% of homeless women are victims of abuse (Zorza, 1991). Among households with children, domestic violence as a cause was reported by thirty-nine percent of cities responding (US Conference of Mayors, 2007).

Sexual abuse

Childhood sexual abuse appears to be a risk factor for adult homelessness (Cauce et al., 2000; McChesney, 1995; Nyamathi et al., 2001; Toro et. al., 1995). The US Department of Health and Human Services (1997) found that 17% of runaway and homeless youth were victims of sex abuse. Research indicates that each year, thousands of children run away from home to escape sexual abuse (Johnson, Rew, & Sternglanz, 2006). Even children placed in foster care may be at risk for running away to homelessness in that the system often fails to provide therapy to deal with the effects of sexual abuse (Roman & Wolfe, 1997). The trauma of sex abuse may well impact school performance, mental health, and other functioning that results in increased vulnerability. Homeless women in particular report a high incidence of childhood sexual abuse. (Bassuk et al., 1997). These childhood experiences are risk factors contributing to impaired healthy development and subsequent difficulties in social and psychological functioning (Molnar, Buka, & Kessler, 2001). In essence, adverse childhood experiences and trauma are strong risk factors for adolescent and adult homelessness (Herman, Susser, Struening, & Link, 1997; Rew, Fouladi, & Yockey, 2002).

Maltreatment

Maltreatment may include sexual abuse, as was discussed earlier. However, maltreatment may be viewed in a broader sense, including physical and emotional abuse, neglect, family conflict, and failure to deal with trauma experiences.

Maltreatment and victimization are risk factors, especially for adolescent homelessness (Cauce, Tyler, & Whitbeck, 2004). Childhood maltreatment is a risk factor for homelessness and housing instability in that it may lead to a cycle of adult victimization and maltreatment by spouses or other relationships (Stein, Leslie, & Nyamathi, 2002). In essence, maltreatment disrupts normal development creating vulnerability. There is evidence that maltreatment may be linked to post-traumatic stress disorder and later homelessness (Gwadz, Nish, Leonard, & Strauss, 2007). Interestingly, there may be gender differences with females more vulnerable. As noted earlier, family maltreatment has been linked to adolescent runaways and flight into homelessness (Gwadz et al., 2007; Kimball & Golding, 2004).

Incarceration

Increasing attention is being given to the relationships between incarceration and homelessness. Homelessness and incarceration are interacting factors in that homelessness increases the risk for incarceration, and the individual being released faces a greater risk of becoming homelessness (Cho, 2008). Research has been consistent in finding that persons from lower socioeconomic levels or minority status are at greater risk for arrest and incarceration (Greenberg & Rosenheck, 2008). Incarceration of these individuals, who have limited resources and social support, tends to isolate them from their communities, thus reducing ability to reestablish or reintegrate into the community (Rosenheck et al., 1998). Six percent of the persons in the 2007 AHAR data had been incarcerated in prison, jail, or detention facilities the evening before entering a shelter. Incarceration as a risk factor is not limited to the United States. A study of more than 22,513 homeless individuals seeking medical services in Moscow found that 30.3% reported being former convicts (Gutov & Nikiforov, 2004). Incarceration increases risks as it interacts with various other factors, such as stigma, loss of skills, discrimination, and often ineligibility for housing or other services.

Being homeless and mentally ill heightens the risk for incarceration (Cho, 2008; Greenberg & Rosenheck, 2008; McNiel, Binder, & Robinson, 2005). The homeless mentally ill have become criminalized and, in a sense, jails have become today's asylums. Even if not initially homeless, the mentally ill person who goes to jail has a significantly increased future risk of housing loss and community disenfranchisement hampering reintegration (Hartwell, 2003; HCH Clinician's Network, 2004; Soloman & Draine, 1995).

Mental illness

The extent of mental illness among the homeless and its role in causation is a source of contention. Studies suggest a range of mental illness from approximately 16% to more than one-third of homeless persons suffering from a severe and persistent mental illness (Federal Task Force on Homelessness and Severe Mental Illness, 1992; US Conference of Majors, 2005; National Resource and Training Center on Homelessness and Mental Illness, 2003). A study of patients discharged from a state hospital was conducted at 13 month and at 3 months after discharge (Belcher, 1988). At the end of 3 months, approximately 35% had become homeless, with more than one-half classified as “aimless wanderers” without any specific plans. Complicating the incidence of mental illness is the number dually diagnosed or with co-occurring disorders such as substance abuse. Being homeless and mentally ill and abusing substances increases the risk for incarceration in criminal justice facilities and subsequently increases the risk of stigmatization and loss of competitive skills for achieving housing and employment.

Domestic violence

Considerable evidence indicates that domestic violence is a primary cause of homelessness, especially for women (ACLU, 2004; US Conference of Mayors, 2007; Zorza, 1991; Homes for the Homeless, 2000). A study sponsored by the Conference of Mayors in 2005 found that 50% of cities surveyed identified domestic violence as a primary cause of homelessness (United States Conference of Mayors, 2005). Ninety-two percent of homeless women report having experienced physical or sexual abuse at some time in their lives (Browne, 1998). Those on welfare seem particularly vulnerable to abuse and subsequent homelessness (Institute for Women's Policy Research, 1997). Domestic violence also involves children (Homes for the Homeless Survey, 2000; Institute for Children and Poverty, 2001), and one study found that 61% of homeless girls and 19% of homeless boys had experienced sexual abuse (Benson & Fox, 2004).

Though leaving the home to escape abuse represents a solution to one problem, the lack of employment and the absence of affordable housing often results in vulnerability for homelessness. The victims of domestic violence are often further challenged by difficulty in finding apartments owing to poor credit and employment histories attributable to the domestic violence (Anti-Discrimination Center of Metro New York, 2005). Discrimination against victims may occur as landlords adopt policies to evict tenants when violence occurs (Tolman et al., 2001). Likewise, some landlords may be reluctant to rent to women who have experienced domestic violence (ACLU, 2004). In a sense, the abused woman is a primary victim of domestic violence but then experiences secondary victimization as she is denied housing and employment because of the history of abuse.

Health status

Health problems such as mental illness and substance abuse are well recognized as risk factors for homelessness. Additionally, other medical problems and infirmities often represent risk factors for homelessness (Rosenheck et al., 1998). Studies of health problems in the homeless population indicate that 46% have chronic health problems such as arthritis, high blood pressure, diabetes, and cancer; 26% report acute infections conditions such as bronchitis, pneumonia, and tuberculosis. Various other conditions including HIV/AIDS are present (Kushel, Vittinghoff, & Ha, 2001; Schanzer, Dominguez, Shrout, & Caton, 2007; Burt et al., 1999).

Health problems in themselves often create stress and anxiety and a sense of vulnerability, which affects one's psychological functioning, further heightening the risk for homelessness. (Rogers, 2008). Just as lack of health care can be identified as a risk factor, medical illness and disability may result in loss of employment and income and bankruptcy (Himmelstein, Warren, & Woolhandler, 2005). For individuals and families struggling to barely meet daily living expenses, a medical crisis can deplete financial resources and push them into homelessness (NHCHC, 2008).

Education

Given the increased requirement for technical and education competence to be self-sufficient, it is logical to assume that poor education is a risk factor for homelessness. In the study “Homelessness: Programs and the People They Serve,” 53% of the parent clients in homeless families had less than a high school education (Urban Institute, 1999). Similarly, in a separate study, fewer than half of young homeless individuals were high school graduates (Burt, Aron, Lee, & Valente, 2001). The percentage completing high school may be misleading as proficiency levels in basics, such as reading and math, may be lower than the actual grade completed (Nooe, 1994).

Substance abuse

Habitual heavy substance use is often cited as a major contribution to homelessness. Compared with the general population, adult homeless persons have a much higher rate of substance abuse (Barber, 1994; Lehman & Cordray, 1993; NCH, 2007; Federal Task Force on Homelessness and Severe Mental Illness, 1992; Tam, Zlotnick, & Robertson, 2003). The presumption that first episodes of homelessness are frequently the result of alcohol abuse was supported by a study of 303 homeless individuals and people at risk of homelessness in Cook County, IL, finding that substance abuse was highly associated with first episodes of homelessness (Johnson, Freels, Parsons, & Vangeest, 1997). Additionally, these findings suggested a multi-directional model in which substance abuse is both a precursor and consequence of homelessness. In a sample of homeless persons at a city soup kitchen, 75% had used drugs in the preceding month (Magura, Nwakeze, Rosenblum, & Joseph, 2000). Single homeless men are especially likely to have histories of substance abuse (Toro et al., 1995). However, care should be taken in interpreting the statistical relationship since studies may over represent shelter users and single men (NCH, 2005). The relationship between homelessness and substance abuse is complex and involves interaction of many of the factors identified in Figure 1. For example, those who are addicted may be impacted by the decrease in SROs. Likewise, the lack of health insurance becomes a barrier in dealing with addiction. Changes reducing eligibility for SSI and Social Security Disability Insurance (SSDI), based on chronic substance abuse, further increase the risk for homelessness. The SSI policy change in 1996 was especially devastating to persons with addictive disorders. These changes resulted in denying SSI and SSDI and disability benefits, including denial of Medicaid eligibility, to persons whose addictions were defined as contributing to their disability status. For many persons, the loss of disability benefits resulted in the subsequent loss of housing (National Health Care for the Homeless Council, 1997, 2005). Similarly, policy changes that result in persons convicted of drug abuse/sale being barred from public housing create additional dilemmas.

In addition to lack of available, appropriate housing, there is also the issue of treatment resources. Many homeless individuals are dually disordered, suffering from both a major mental illness and substance abuse (Barber, 1994; Task Force on Homelessness and Severe Mental Illness, 1992). These dually disordered individuals frequently fall between the cracks because neither mental health nor substance abuse treatment facilities provide comprehensive services. Though substance abuse contributes to the lack of funds and eligibility for housing, it may also increase family conflict leading to family unwillingness to allow individuals to reside in the home.

Although addiction frequently is cited as a contributing factor, it must be considered within the context of housing, employment, and treatment availability (Oakley & Dennis, 1996). Public attitudes around addiction and the earlier cited stereotypes regarding homelessness and substance use may well impact the lack of programs to address these problems and the elimination of disability benefits.

Minority status

Minority status as a risk factor is based on the fact that blacks and other minorities are more likely to be poor, thus at greater risk for homelessness. Perhaps the most telling factor is that in 2006, approximately 21% of Hispanics and 24% of blacks were in poverty (US Census Bureau News, 2007). Therefore, minority status may not only increase the risk for homelessness but increase the barriers to escaping homelessness (Task Force on Homelessness and Severe Mental Illness, 1992; Rosenheck et al., 1998). Homelessness disproportionately impacts minorities constituting about 59% of the sheltered population compared with representing 31% of the US population (USDHUD, 2007). The homeless population is disproportionately black, approximately 45%, compared to the general population wherein blacks are about 12% (US Census Bureau, 2003). The loss of jobs in the inner city, housing segregation, and other structural factors are obviously intertwined with poverty to increase the risk for minorities. There may be racial differences among the causes of homelessness in that whites report more internal causes, such as substance abuse and mental illness, compared to non-whites reporting more external factors such as low income and unemployment (North & Smith, 1994).

Studies on homeless whose minority status is based on sexual orientation are scarce. However, studies examining the sexual orientation of homeless youth suggest rates ranging from 6% to 40% (Cochran, Stewart, Ginzler, & Cauce, 2002). The National Gay and Lesbian Task Force suggests that between 20% and 40% of all homeless youth identify as lesbian, gay, bisexual, or transgender. The report goes on to say that family conflict over a youth's sexual orientation often leads to exclusion from the home and subsequent homelessness (Ray, 2006). Shelters and other care facilities may also react to a youth's sexual orientation in a manner that results in leaving and becoming homeless (Hyde, 2005).

Military service

Approximately 41% of homeless men are veterans compared to 34% of male veterans in the general population (National Coalition for Homeless Veterans, 2008; National Coalition for the Homeless, 2008; Rosenheck, Frisman & Chung, 1994). Gamache, Rosenheck, & Tessler (2003) found that women veterans are 3.6 times more likely to be homeless than non-veteran women. However, Rosenheck asserts that homelessness among veterans is not clearly related to military experience, but the same factors—poverty, lack of housing, alcohol and drugs—that contribute to homelessness among non-veterans are significant risk factors for homelessness among veterans (Rosenheck, Leda, Frisman, Lam, & Chung, 1996). Military service may be associated with other risk factors such as substance abuse and criminal behavior in becoming homeless. In a study of 188 homeless veterans, Benda, Rodell, and Rodell (2003) found that among those who were substance abusers, 41% reported committing crimes in the past year. Drug and alcohol abuse, lower levels of education, unemployment, physical or sexual abuse before age 18, and habitation with a substance abuser elevated the likelihood of homeless veterans committing crimes. The odds of committing a crime were decreased by resilience, ego integrity, and self-efficacy, all of which are individual factors.

Individual and Social Outcomes

As is evident from the preceding discussion, multiple biopsychosocial risk factors can interact to produce homelessness. Once homeless, individuals and families are vulnerable to a number of harmful individual and social outcomes. Figure 1 displays these individual and social outcomes in the context of our model of the ecology of homelessness. It is to these outcomes of homelessness that the discussion now turns.

Individual outcomes

Health impairment

Health and health care reflect the interrelationship between personal and structural factors contributing to homelessness and may represent both a risk factor for homelessness and a result of homelessness. Chronic and acute health problems are frequent among the homeless (NHCHC, 2005). At the same time, homeless individuals are often uninsured and lack access to basic health care (O'Connell, Lozier, & Gingles, 1997) Illness or disability often results in lost employment and eviction followed by homelessness and overwhelming challenges to regaining stability.

Kushel, Perry, Bangsberg, Clark, and Moss (2002) surveyed 2,578 marginally housed and homeless persons in San Francisco regarding their use of hospital emergency department services in the prior year. Slightly more than 40% visited the emergency department at least once during the year. Strikingly, 7.9% respondents were responsible for 54.4% of all visits for the group. High use rates were associated with substance abuse, physical and mental illness, victimization, arrests, and less stable housing.

Homeless smokers (n = 107) were compared to 491 non-homeless smokers in a study of the smoking habits of homeless smokers (Butler et al., 2002). Homeless smokers were found to smoke more cigarettes per day, to have begun smoking at a younger age, to have smoked for longer period of time, had a higher level of depression, and were more likely to use recreational drugs.

Homeless individuals are at greater risk of latent and active TB secondary to time spent in homeless shelters and other homeless services delivery facilities. In a study of 415 homeless persons in Los Angeles, Nyamathi et al. (2004) found a notable lack of knowledge regarding the risk factors associated with TB infection and modes of transmission. Injecting drug users and Latinos evidenced lower rates of TB knowledge.

Homeless children in particular suffer health problems, often resulting from hunger and poor nutrition. Development delays are frequent among children. Psychological problems such as anxiety, depression and behavioral problems are elevated (Rafferty & Shinn, 1991). Various childhood diseases, delays in immunizations, and lack of health care are much greater among homeless children than the housed poor and national norms (Wright, 1991).

Substance abuse

The relationship between substance disorders and homelessness is complex. It is likely that for many individuals, the risk of homelessness is increased by substance abuse/dependence, but substance use may be accelerated by homelessness. Among cities surveyed by the United States Conference of Mayors (2008), 68% reported substance abuse as one of the leading cause of homelessness. Many persons become addicted while homeless (NCH, 2007). A critical factor is that homelessness may create additional barriers to treatment for substance disorders, including lack of insurance, lack of supportive resources, and general inaccessibility of programs (NASADAD, 2007).

Mental illness

Burt (1992) reports that mental illness and/or chemical dependency problems affect more than 50% of the US homeless population. As was noted in earlier discussion, for persons with mental illness and substance abuse, homelessness has a detrimental effect and, like any other crisis or trauma, may “catalyze and/or exacerbate mental illness producing disorder where previously if did not exist” (Koegel & Burnam, 1992). Banyard and Graham-Bermann (1998) found that homeless mothers had more depression and used avoidant coping strategies more than housed mothers. However, it may well be that depression and avoidance are a consequence rather than cause of homelessness.

Social isolation

Regardless of the factors involved, the availability of social support whether from friends, relatives, or agencies appears to influence both risks for and recovery from homelessness. Kingree et al. (1999), for example, found that low levels of support from friends were associated with homelessness after completion of a substance abuse treatment program. Similarly, personal crisis such as divorce and widowhood remove support systems and may make individuals more vulnerable to homelessness.

Impaired education

Poverty and homeless have a devastating impact on educational achievement. In both the housed poor and the homeless; children in particular experience an increased risk of inability to succeed in school or community environments (Ziesemer, Marcoux, & Maxwell, 1994). Homeless children often experience difficulty accessing and staying in school. Even when enrolled, instability and shelter-living are not conducive to learning. Compared to their peers, homeless children are likely to experience educational underachievement and often are required to repeat grades (Rafferty & Shinn, 1991; Institute for Children and Poverty, 2003). Impaired education in childhood likely has long-term consequences, increasing the risk for adult poverty and homelessness.

Sexual abuse

There is increasing evidence that the incidence of childhood sexual abuse in the histories of chronically homeless women and clinical populations is much higher than the general population (Bassuk, Dawson, Perloff, & Weinreb, 2001; Kushel, Evans, Perry, Robertson, & Moss, 2003; Molnar et al., 2001). After becoming homeless, the risk of sexual assault may be heightened, with some, such as, transgendered persons, being particularly vulnerable (Kushel et al., 2003). The limited studies available suggest that more than 10% of homeless women have been raped, many multiple times, during the past year (Wenzel, Leake, & Gelberg, 2000; Heslin, 2004). In addition to the trauma of sexual assault, many of these homeless individuals encounter barriers to services, including social stigma, lack of information or access, and lack of insurance and transportation (Pennsylvania Coalition Against Rape [PCAR], 2006).

Maltreatment

Earlier, maltreatment was recognized as a contributing or risk factor for homelessness. Once homeless, the incidence of maltreatment may be no less severe. Perhaps one of the most impacting outcomes is being stigmatized and blamed for being homeless. Thus, in addition to the hardships of homelessness, these individuals are often labeled and ostracized by their communities (Phelan et al., 1997).

Advocates identify a trend toward “criminalization of homelessness.” In other words, using the criminal justice system to respond to homelessness rather than seeking more human solutions (NLCHP, 2006). The NCH and Saelinger (2006), for example, cite the trends of revitalization of downtowns and compassion fatigue toward the poor as underlying this shift.

In addition to stigma and criminalization, homeless persons face increased risk of physical maltreatment. Homelessness places women at higher risk of assault and rape (Heslin, 2004; PCAR, 2006). Homeless persons are victimized disproportionately as compared to the domiciles population (Lee & Schreck, 2005). According to the NCH, from 1999 through 2006, there were 774 acts of violence, resulting in 217 murders of homeless people and 557 victims of non-lethal violence in 200 cities from 44 states and Puerto Rico (2008). These attacks seem to be increasing, suggesting that these figures may be conservative in that studies often report as many as two thirds of homeless persons experience victimization (Mallory, 2002).

Criminal activity

Just as homeless persons tend to be victimized more than the general domiciled populations, they also have disproportionately higher incarcerations rates. Homeless jail inmates comprise more than 15% of jail inmates. Compared to other inmates, they are more likely to have committed a property crime, have a history of justice system involvement, suffer from mental illness and substance abuse, and be older, unemployed, and single (Greenberg & Rosenheck, 2008). Given these characteristics, it is logical to assume that they are more disadvantaged and face greater barriers to community reintegration.

Though the foregoing suggests that homeless persons are disproportionately incarcerated, care should be taken before concluding that homelessness has an outcome of increased criminal activity. The passage of anti-nuisance laws in recent years has expanded the definition of criminal activity and increases the risk of arrest and incarceration. Ordinances that prohibit panhandling, camping, and eating and loitering in public spaces increases the risk of arrest for activities that may be part of survival for homeless persons (NCH & NLCH, 2006). Though the research is limited, it may well be that criminal activity by homeless persons may be higher for larceny offenses and ordinance violations but not significantly higher for violent, property and sexual offenses. (Polczynski, 2007). However, in an earlier study, Martell (1991) found the mentally ill homeless represented 50% of all admissions to a New York City maximum security hospital. This study suggested a higher base rate of violent criminal behavior in this population and the existence of a subgroup that might pose a threat to public safety secondary to violent behavior.

Criminal victimization

Homeless people are victimized disproportionately compared to the domiciled population entailing various forms of victimizations (Lee and Schreck, 2005). Sexual and physical assault are particularly common experiences among the homeless. A study found that 32% of women, 27% of men, and 38% of transgendered homeless persons reported physical or sexual assault in the previous year (Kushel et al., 2003). A California study found that 66% had been victimized in the previous year, with many reporting multiple victimization; assault and robbery were the most frequent reported by approximately three-fourths of the respondents (Mallory, 2002).

It has been pointed out that dimensions of marginality—disaffiliation, health problems, and traumatic events, for example—significantly increase the odds of being victimized (Lee & Schreck, 2005). Homeless persons with severe mental illness or other disabilities may be particularly vulnerable to violent criminal victimization, with some studies indicating a two-and-a-half times greater risk (Hiday et al., 1999). The NCH (2006) details the increasing rate of violent crimes in its report, “Hate, Violence and Death on Main Street USA: A Report on Hate Crimes and Violence Against People Experiencing Homelessness 2006.” Perhaps most disturbing is that many of these crimes are committed by adolescents and young adults for no motive other than boredom. It is the NCH's position that these constitute hate crimes, even though not currently included in federal law.

Job loss

Job loss and unemployment were identified earlier as risk factors for homelessness. The loss of long-term jobs and involuntary job loss has increased in recent years (Mishel et al., 1999). It is not surprising that being homeless may further diminish one's chances of employment, both in terms of opportunities and as prolonged idleness may result in loss of skill and work habits, responsibility, and commitment to employment. Those with limited education, experience, or skills are especially at risk for prolonged homelessness. In addition to struggling with basic issues of survival such as food and shelter, the homeless worker often faces barriers such as lack of transportation, childcare and treatment for health problems. (Long, Rio, & Rosen, 2007; US Department of Labor, 1994). Even the lack of appropriate clothing for employment may pose a risk for maintaining or finding employment.

Self-harm

Service providers offer many anecdotal accounts of suicidal ideation and attempts by homeless persons, but the research is sparse. The frequent characteristics among the homeless—social isolation, mental illness, substance abuse, and poverty along with previous attempts of self-harm—represent high suicidal risk factors (Christensen & Garces, 2006). A study of homeless and runaway adolescents found that more than one-half had experiences suicidal ideation and more than one-fourth had made an attempt during the previous year (Yoder, Hoyt, & Whitbeck, 1998).

Suicidal ideation and attempts among mentally ill homeless persons are especially high. There is some evidence that young middle-aged persons are at greatest risk of suicidal behavior, a pattern quite different from the risk factors in non-homeless populations of older individuals (Prigerson, DeSai, Liu-Mares, & Rosenheck, 2003). Regardless of the age group or health status, homeless often result in a sense of extreme hopelessness, which often underlies self-harm.

Death

One of the tragic outcomes of homelessness for some is death. Multiple studies have found elevated morality rates among homeless individuals ranging from 2 to 8.3 times higher than the general population, depending on gender, age group, and setting (Hibbs et al., 1994; Hwang, Orav, O'Connell, Lebow, & Brennan, 1997; Barrow, Herman, Cordova, & Struening, 1999; Hwang, 2000). In a study of 1981 women, Cheung and Hwang (2004) found morality rates for homeless women 10 times higher than for women in the general population. When they examined mortality data for homeless women in seven cities, they found women younger than 45 years had a mortality rate near or equal to men of a similar age. The most common cause of death was HIV/AIDS and drug overdose.

Social outcomes

Poverty

Poverty and homelessness are intertwined. In a sense, homelessness represents the “poorest of poor.” In 2002, people below the official thresholds numbered 34.6 million, a figure 1.7 higher than the 32.9 million in poverty in 2001 (Proctor & Dalaker, 2003). An earlier study of single homeless individuals found an average income approximately 51% of the 1996 federal poverty level (US Interagency Council on the Homeless, 1999). The declining value of public assistance and shifts in welfare policy have resulted in fewer resources and stricter guidelines for subsidies and services (Berger & Tremblay, 1999; NCH, 2009; Dunlap & Fogel, 1998). These trends have increased the poverty level for homeless persons and make escape more difficult. Resources such as AFDC were important in preventing homelessness, but more exclusionary guidelines have negatively impacted preventing and escaping of homelessness.

Public safety

McNeil et al. (2005) found in a study of 12,934 individuals incarcerated in the San Francisco County Jail that 16% were homeless, with 30% of the homeless inmates having a severe psychiatric disorder. The homeless were significantly more likely to (1) receive a diagnosis of a co-occurring disorder, (2) be charged with a felony, but (3) not be charged with a violent crime. Inmates who were both homeless and diagnosed with a serve mental illness and substance abuse were more likely to have multiple incarcerations than inmates who were neither homeless nor dually diagnosed. As Greenberg and Rosenheck (2008) found, homeless inmates are more likely to have been charged with violent offenses, thus raising questions about public safety. The previously noted factors of untreated mental illness, substance abuse, disaffiliation, and lack of social support may increase issues of recidivism and subsequently public safety (Martell, 1991; Polczynski, 2007).

Property crime

There is often a public perception that homeless persons frequently commit property crimes (Tepper, 2006). As noted, in comparison with other jail inmates, the inmate who has been homeless is more likely to be incarcerated for a property crime. The homeless inmate is more likely to have a history of criminal justice involvement, have mental health and substance abuse problems, and be unemployed and less educated (Greenberg & Rosenheck, 2008). However, findings have been consistent that the majority of arrests, around 80%, are for substance abuse and offenses such as trespassing. However, there are mixed results. There is evidence that the homeless arrest rate for serious crime exceeds the arrest rates for those with permanent addresses. The rates for both property crimes and violent crimes have been significantly higher in some studies (Snow & Anderson, 1993). At the same time, other studies suggest that violent and property crimes are not correlated in areas close to homeless services (Polczynski, 2007).

Prostitution

Homeless women and runaway youth are especially vulnerable to falling into prostitution. Studies of homeless youth involved in prostitution range from ten to fifty percent. (Greenblatt & Robertson 1993). A Chicago study found that 50% of women involved in prostitution had been homeless (Mueller, 2005). Prostitution is frequently associated with drug addition (McClanahan, McClelland, Abram, & Teplin, 1999; Silbert, Pines, & Lynch, 1982). However, survival sex—the exchange of sex for money to secure food, shelter, and other basic needs—is likely intertwined with other factors.

Earlier, childhood sex abuse was discussed as a risk factor for homelessness. Similarly, childhood sex abuse has been identified as an antecedent to prostitution (Simmons & Witbeck, 1991; Seng, 1989; McClanahan et al., 1999; Bagley & Young, 1987). Most likely the path is not linear but represents the interaction of abuse, runaway behavior, homelessness, substance abuse, and prostitution. Unfortunately, these factors create additional barriers to escaping homelessness and negative social consequences.

Street violence

Street violence is interlinked with homelessness. Earlier discussion recognized the rates of sexual abuse and other types of criminal victimization. It takes many forms ranging from assault by homeless persons on one another to thrill assaults and killings by teens and even to abuse by police and medical responders. Homeless persons experience physical and sexual violence much more frequently than housed persons (Street Health, 2007) The NCH publishes annual reports on hate crimes and violence, noting “In 2006 homeless individuals in America faced another year of brutality that ranged from assault to killings” (NCH, 2007). A study in Toronto, Canada, found a rate of physical assaults 35 times higher than the housed population (Street Health, 2007). Persons living outside, especially those with mental illness or substance abuse are particularly vulnerable to violence (NCH, 2007; Hiday et al., 1999). Street violence is pervasive and wide-ranging, with approximately 10% of homeless persons report being assaulted by police (Street Health, 2007; Hwang, 2004).

Public inebriation

Habitual heavy substance abuse is not uncommon among the homeless, increasing the risk for public inebriation and subsequent arrest (NCH, 2007; US Conference of Mayors, 2005, 2007; Toro et al., 1995). In some community studies, closer to 60% of homeless adults had been arrested for public inebriation within the past 3 years (Nooe, 2006). Public inebriation and incarceration may become cyclic, creating additional barriers to reintegration (Hartwell, 2003; NHCHC, 2005). The lack of health insurance, the limited number of resources for treating dually diagnosed individuals, and lack of support services are additional elements (Hartwell, 2003; HCH, 2007). Likewise, policy changes that result in persons convicted of drug abuse/sale being barred from public housing have created additional dilemmas. Another social outcome is the cost of public inebriation. Dunford found that 15 randomly selected chronic alcoholics amassed costs of $1.5 million for emergency department care during an 18-month period (Dunford et al., 2006).

Panhandling

The president of the Denver Metro Convention and Visitors Bureau is quoted as saying, “Panhandling and homelessness is still the number 1 negative comment from our tourists and conventioneers” (Steers, 2007). Whether accurate, there appears to be a widely held public view that homelessness and panhandling are closely related (Lee & Farrell, 2003; Rimer, 1989). However, studies offer different estimates of the number who panhandle, ranging from around 5% to more than a third of homeless people (Burt et al., 1999; Rossi, 1989; Snow, Anderson, Quist, & Cress, 1996). Likewise, results are mixed in addressing questions about panhandling being necessary for survival and the productivity of it (Lankenau, 1999; Snow et al., 1996; Stark, 1992). The literature is fairly consistent in suggesting that a person panhandling is more likely to be male and minority, and increases with duration of homelessness. (Lankenan, 1999; Stark, 1992). Lee and Farrell, 2003, reviewing the literature, indicate that panhandlers are generally “more isolated, troubled and disadvantaged,” having more extensive histories of mental illness, substance abuse and criminal records than non-panhandling homeless persons.

Though information about homelessness and panhandling is limited, it is apparent that cities throughout the United States are enacting ordinances prohibiting or severely restricting panhandling (NCH & NLCHP, 2006). Thus, panhandling may be received as a social outcome of homelessness and stimulating social outcomes of ordinances and regulations.

Strained health services

The increased risk for health problems due to homelessness is well documented (Schanzer et al., 2007; Plumb, 1997). A Baltimore study found that homeless adults have an average eight to nine concurrent medical illnesses (Breakey et al., 1989). Rates of substance abuse and mental illness among the homeless are likely double those of other low-income patients (Fisher & Breakey, 1991). The nature of homelessness—extreme poverty, social isolation, and lack of support networks—suggests greater risk for poor health status.

Homeless individuals use more impatient and emergency department services rather than outpatient or clinic services (Salit, Kuhn, Hartz, Vu, & Mosso, 1998; Folsom et al., 2005). The Centers for Disease Control National Health Survey found that approximately one-third of homeless individuals, as compared to 1% of the general population, used emergency facilities rather than clinics or private physicians for primary care.

The high incidence of substance abuse and mental illness among homeless persons frequently seen in emergency departments results in strained health services in terms of treatment and financial resources (Dunford, et al., 2006). When admitted, the homeless patient generally stays approximately 4 days longer than other comparable patients (Salit et al., 1998). Many of these individuals will be discharged back into homelessness, increasing the risk for continued illness or relapse. As noted earlier, most homeless persons are uninsured and, as need for treatment increases, the health system is further strained (O'Connell, Lozier, & Gingles, 1997). An area of concern is the estimation that 6% to 27% of the homeless population in the United States are HIV positive (Kim, Kertesz, Horton, Tibbetts, & Samet, 2006).

Community philanthropy

Although they may experience “compassion fatigue,” the public seems willing to support efforts to solve homelessness (Link et al., 1995). An earlier study by Toro and McDonnell (1992) found that more than 50% of respondents were willing to pay more taxes to help homeless persons. Though there is an expressed willingness to give, a number of factors my influence actual giving. Morgan, Goddard, and Givens (1997) examined the relationship between expressed willingness to assist the homeless and an individual's level of empathy, religion, household income or political orientation, gender, and race. In a study of 204 undergraduates, the authors found empathy and expressed religiosity strong predictors of willingness to assist the homeless.

Religious organizations play a major role in philanthropy. Cnaan and Handy (2000) examined the social services provided by religious congregations in Ontario and the United States. They found that 56.5% of Canadian congregations provided shelter for men compared to 28.7% of U.S. congregations; 54.3% of Canadian congregations provided shelter for women/children as compared to 28.3% of U.S. congregations.

Community philanthropy directed toward programs for the homeless also can be considered in terms of foundation giving and support from faith-based organizations. Historically, only about 1% of foundation giving has been for homeless giving; however, there are renewed efforts to involve foundations (National Center on Family Homelessness, 2008). Overall, churches and other faith-based organizations have played a more prominent role in addressing homelessness. The 1996 National Survey of Homeless Assistance Provides and clients found that the majority of all food programs and one-fourth of all shelters are operated by faith-based organizations (Aron & Sharkey, 2002).

Community discord

Community discord is often a social outcome of the increase n homelessness. Newspapers reflect these conflicts in stories such as “Atlanta Puts Heat on Panhandlers” (USA Today); “Pressed on the Homeless, Subways Impose Rules” (New York Times, 1989); and “Church Sued for Housing Homeless” (Erie Times–News, 2008). Saelinger (2006), summarizing a number of studies, notes that during the 1980s, homelessness was viewed as a critical social issue needing human solutions but, as cities revitalized and compassion fatigue increased, communities turned to anti-nuisance laws. Two major manifestations of community discord are the anti-nuisance ordinances regulating activity in public spaces, including prohibiting sleeping, feeding, and begging (NCH, 2005) and conflict and protest against location of shelters and homes for the homeless (NLCHP, 1997; Lyon-Callo, 2001).

The NCH report “Illegal to be Homeless: The Criminalization of Homelessness in the United States” (2004) documents the organization's view that “through the passage of possibility unconstitutional laws, the ‘selective enforcement’ of existing laws, arbitrary police practices and discriminatory public regulations, people experiencing homelessness face overwhelming hardships in addition to their daily struggle for survival.” An indication of the level of discord is reflected in a New York Post article suggesting that the American Civil Liberties Union is an instigator:

What the homeless industry really wants is total exemption from the law for street vagrants so that they can remain publicly visible until the final thieves of alcoholism and schizophrenia drive them to the hospital or the grave. It is the enforcement of the laws—period—that infuriates these advocates, not their alleged “selective” enforcement. (MacDonald, 2002)

Community discord is often heated regarding the location of shelters and housing of homeless persons as reflected in “Not in My Backyard or NIMBY” (NLCHP, 1997). Though advocates often cite accessibility to resources, transportation, and centralization of services, those opposing cite decreased property values and increased crime rates (NLCHP, 1997).

The underlying reasons for community discord are complex, often involving an interplay between representations of homeless people and historical, class, and power dynamics within communities (Lyon-Callo, 2001). Takahashi (1998) points out that interconnections such as HIV/AIDS and homelessness add an additional dynamic that many be represented in NIMBY as a product of the changing social construct of stigma. As noted earlier, homelessness was a major issue for resolution in the 1980s, but Americans often grow impatient when solutions are slow in coming, perhaps reflecting “compassion fatigue” (Link et al., 1995).

Temporal Dimension

Homelessness is a phenomenon of variable duration. Individuals and families may experience homelessness over the course of a single evening or extending multiple years. The categorization of temporal dimension of homelessness presents a significant definitional challenge. Commonly used terms describing the duration of homelessness include “first time,” “short-term,” “situational,” “transitional,” “episodic,” and “chronic.” Each of these terms has associated limitations, though “first time” conveys the most obvious specificity. The period of time denoting “short-term” homelessness is not well defined in the research literature. Perhaps more concretely, though researchers might legitimately view a 2-week period of homelessness as “short-term,” to a family experiencing 2 weeks of living in an emergency shelter, “short-term” might seem interminable. “Transitional” was defined by Kuhn and Culhane (1998) as one stay in an emergency shelter for a “short-period” (p. 207). Alternatively, “transitional” housing programs typically provide residents shelter up to 24 months. Episodic homelessness is defined by frequent short periods of homelessness alternating with periods of time spent in temporary housing or institutions such as jails, detoxification centers, or psychiatric facilities (Farr, Koegel, & Burnam, 1986; Kuhn & Culhane).

The US federal government defines a “chronically homeless” person as “an unaccompanied homeless individual with a disabling condition who has either been continuously homeless for a year or more, or has had at least four episodes of homelessness in the past three years” (Notice of Funding Available for the Collaborative Initiative to Help End Chronic Homelessness, 2003). This definition conflates duration with condition. A person who has been “continuously homeless” for 1 year or experienced “homeless episodes during the last three years” could be considered chronically homeless and must also have a “disabling condition.” Disabling conditions can include severe and persistent mental illness, severe and persistent alcohol and/or drug abuse problems, and HIV/AIDS.

It is noteworthy that the absence of a permanent residence or repeated periods of homelessness is insufficient to meet criteria for chronic homelessness. The US federal definition requires linkage to a plausible causal factor, a disabling condition. Strikingly, the allowable plausible causal factors center around individual disability, eschewing socioeconomic engendered causality such as unaffordable housing, low wages, and job loss. Further, the definition excludes children who are homeless with their parents, unaccompanied individuals without disabilities, and unaccompanied individuals who elect not to declare a disability.

Some researchers have posited a typology of homelessness that links duration or number of episodes of homelessness to groups or types of homeless individuals (Caton et al., 2005; Kuhn & Culhane, 1998). For instance, Kuhn and Culhane assert that transitionally homeless individuals are younger and typically have fewer problems associated with mental health, substance abuse, or medical conditions. Others have noted greater similarity than differences between the first-time homeless and those individuals with multiple episodes of homelessness (Goering, Tolomiczenko, Sheldon, Boydell, & Wasylenki, 2002).

The temporal dimension of homelessness is represented in our ecological model of homelessness (see Figure 1) with the categories First Time, Episodic, and Chronic, as these three terms are the most definitionally distinct. Biopsychosocial risk factors associated with homelessness can result in any one of these three temporal categories. It is important to recognize that every first-time homeless individual is at some risk of becoming episodically or chronically homeless. Individuals may move from being episodically homeless to chronically homeless by having a disabling condition and four or more episodes of homelessness in the previous 3 years. Conversely, it is possible that an individual previously categorized as chronically homeless could after a period of housing stability again become episodically homeless. This discussion now turns from the temporal dimension of homelessness to its locus and residential alternatives.

Housing Status

The ecological model of homelessness in Figure 1 depicts a continuum of living arrangements experienced by homeless individuals and families. The arrows connecting these housing arrangements are intended to represent the real and potential pathways of transition between these housing categories. The arrows linking housing status categories to individuals and social outcomes represent the array of consequences associated with homelessness. This discussion begins with perhaps the most visible manifestation of homelessness: street dwelling.

Street dwelling

The term street dwelling is used here to denote a spectrum of impermanence or transitory sleeping options including on the streets, in vehicles, and in temporary camps. Sleeping rough is the British term for individuals who eschew staying in homeless emergency shelters for nights on the street (Urban Dictionary, 2008). In the 2007 Annual Homeless Assessment Report (AHAR) to Congress (Khadduri et al., 2007, p. i.), “unsheltered” is the term used for persons “who do not use shelters and are on the streets, in abandoned buildings, or in other places not meant for human habitation.” For many residents of urban centers, street dwelling homeless may be the most familiar or commonly encountered among the wider homeless population. They sleep on the streets, under overpasses, in abandoned buildings, and in public spaces. Street-dwelling homeless traverse urban environments commonly carrying their few possessions, having limited or no access to sanitary facilities, and encountering an array of unsafe and unhealthy conditions (Cousineau, 1997)

O'Connell et al. (2004) conducted a 4-year prospective study of 30 street-dwelling individuals older than the age of 60. Over the course of the study, 30% (9) died, 27% (7) continued to live on the streets, 20% (6) moved into nursing homes, 17% (5) were housed, and 1 was lost to follow-up. The health and safety risks associated with street dwelling are not limited to elderly homeless. Ferguson (2007) lists high-risk survival behaviors of street-dwelling youth including prostitution, panhandling, pornography, drug dealing, and other criminal activities. Health problems of street-dwelling youth include HIV/AIDS and other sexually transmitted diseases, substance abuse and dependence, malnutrition, skin and respiratory inflections, and a host of mental health problems.

The 2007 AHAR study (Khadduri et al., 2007) reports in a point-in-time study conducted in January 2005 that 45% of all homeless individuals were unsheltered. Of that group, 30% were chronically homeless, compared to the 17% rate of chronically homeless found among the sheltered homeless in the same point-in-time study. The authors point out that street counts of the unsheltered homeless may be inflated owing to methodological problems.

Some studies have attempted to differentiate unsheltered homeless persons from those found in emergency shelters and transitional housing. Larsen, Poortinga, and Hurdle (2004) used an unmatched, case control study to compare 85 homeless individuals who used the services of local shelters with 45 individuals who did not use shelters. The homeless individuals who elected to not use shelters were more frequently employed as day laborers, consumed large quantities of alcohol more frequently, and were more likely to be Native-American and to have received court-ordered psychiatric treatment. Lam and Rosenheck (1999) report that unsheltered homeless have a greater probability of being less interested in treatment and are more likely to have psychotic disorders, to be older, to be male, and to be more difficult to engage in case management.

Shared dwelling

The housing category “shared dwelling” refers to an option employed by individuals at risk of becoming homeless or transitioning out of homelessness for some period of time. Individuals using the shared dwelling option are sometimes referred to as couch homeless, precariously housed, doubled up, couch surfers, or the hidden homeless owing to the methodological challenges of finding and counting them. Nonetheless, Hoback & Anderson (2006) estimate that on any night in 2000 there were 4,700,000 couch homeless in the United States. This represents 1.65% of the population. Burt (1996) suggests that the “precariously housed” typically are unable to afford housing and that their shared dwelling arrangement usually lasts less than 60 days. According to the 2007 AHAR study (Khadduri et al., 2007), these individuals are commonly “doubled up” with relatives and friends and are at risk of homelessness.

Emergency shelter

The term emergency shelter refers to shared housing provided to homeless individuals and families. This shelter option is sometimes referred to as a homeless shelter or mission. They are typically found in urban settings and commonly have few restrictions on admission. The allowable length of stay in an emergency shelter depends on shelter policies and funding source restrictions. A Google search of the term homeless shelter length of stay produced documents reporting shelter policies on length of stay ranging from one night to up to 1 year. Wong, Park, and Nemon (2006) report that in a survey of 300 homeless residential programs, 43% of emergency shelters had either no length-of-stay restrictions or had no formal policy on length of stay.

Services provided by emergency shelters range from the basic bed and meals to more rehabilitative services including substance abuse treatment, domestic violence assistance, job placement skills, case management, and housing assistance. Many shelters discharge clients each morning and re-admit them in the late afternoon, forcing clients to spend their days outside the shelter. Other shelters provide around-the-clock shelter, food, and services. Callicutt (2006, p. 169) suggests that homeless shelters, much like jails and prisons, are now part of a “de facto mental health system” that now provides housing to the severely persistently mentally ill who formerly would have received care in long-term psychiatric facilities.

Emergency homeless shelters are operated by city or county agencies, nonprofit groups, churches, and religious organizations supported by local and national church organizations. HUD provides funding for homeless shelters to its emergency shelter grants program. This funding is available to faith-based, secular organizations and local government agencies, though faith-based organizations are required to provide services in a manner free from religious influence (HUD, 2008).

Jail and prison

Jails and prisons across the United States regularly house the homeless and individuals destined for future homelessness. The relationship between homelessness and incarceration appears reciprocal; being homeless increases ones chance of arrest and once incarcerated, the risk of homelessness increases. Burt et al. (1999) found 54% of homeless individuals reported having been incarcerated for 5 days or more in a city or county jail, state or federal prison, or juvenile detention.

McNeil et al. (2005) examined their records of 12,934 individuals incarcerated in the San Francisco County Jail in the course of 6 months in 2000. They found that 16% of those incarcerated were homeless and that 30% of those individuals had a diagnosis of a mental disorder. Co-occurring substance abuse–related disorders were found in 78% of individuals with a severe mental disorder. Individuals with severe mental disorders and co-occurring substance abuse disorders were found to be held in jail longer than inmates charged with similar crimes.

Greenberg and Rosenheck (2008) studied data from a national survey of adults in state and federal prison. They found that the rate of homelessness among inmates was four to six times higher than the general population's estimated rate of homelessness. Strikingly, homeless inmates were more likely to be poor and have a history of trauma, substance abuse, mental health problems, and poor health. Additionally, homeless inmates also were more likely to have a record of prior property and violent offenses.

Metraux and Culhane (2006) matched data from New York City jails and New York State prisons to the homeless shelter records of 7,022 individuals who had spent time in a New York City public shelter over a 2-year study period. They found that 23.1% of the population had been incarcerated in the previous 2 years. Individuals released from jail, 17% of the study group, were more likely to have used shelters more frequently and evidenced a sequential pattern of alternating jail and shelter stays resulting in continuing residential instability. Individuals released from prison constituted 7.7% of the study group. This group was more likely to have a shelter stay within 30 days of release from prison and were less likely to have subsequent shelter stays when compared to individuals released from jail. The authors called for the development of differential approaches to preventing homelessness for individuals released from prison versus those released from jail.

Hospitalization

Inpatient hospitalization of homeless individuals, whether for substance abuse, mental illness, or health reasons, accounts for an expensive but unknown proportion of the time the homeless are sheltered. Moore (2006) examined the 5-year cost of health care and incarceration of 35 duly diagnosed, chronically homeless individuals in Portland, OR. The average annual cost per individual was $42,075. Hospitalizations associated with health care, mental illness, and substance abuse accounted for 90% of the total 5-year expenditure of $7,363,214. The average annual cost per person found by Moore is similar to the findings of Culhane, Metraux, and Hadley (2001). Their study of a similar population in New York City found the approximate annual cost of major services to be $40,500.

Salit et al. (1998) compared 18,864 homeless adults hospitalized in New York City in 1992 and 1993 to 383,986 low-income adults admitted for non-maternity reasons during the same time period. They found that homeless adults remained in the hospital 36% longer, which was partially attributed to the difficulty of finding appropriate discharge placements. In 80.6% of the admissions of homeless adults, substance abuse or mental illness was a primary or secondary diagnosis, double the rate for non-homeless, low-income adults in the study. The authors called for the development of supportive and low-cost housing options.

Culhane, Metraux, and Hadley (2002) found that placement in supportive housing (described further) reduced the number of days spent in inpatient psychiatric hospitals by 61% and public (non-psychiatric) hospitals by 21%. Similarly, Perlman and Parvensky (2006) found an 80% reduction of inpatient hospitalization nights associated with the placement of chronically homeless individuals in supportive housing.

Transitional housing

Transitional housing programs enable homeless individuals and families to move into a more stable housing situation for up to 24 months. During this time period, clients can address issues associated with their prior homelessness including substance abuse, mental health problems, poor credit/rental histories, employment, and other personal and situational impediments to permanent housing. Transitional housing programs are typically smaller and more focused on client rehabilitation, behavior change, and resource linkage than emergency shelters (Levinson, 2004). Crook (2001) found that residents reported more positive experiences in transitional housing programs that evidenced greater indigenous participatory leadership, decreased bureaucratic control, and higher levels of personalized interaction with residents.

HUD funds transitional housing programs through its Supportive Housing Program (HUD, 2008). These programs are designed to facilitate the movement of previously homeless clients to permanent housing by providing support services including home furnishings, childcare, and job training. Wong et al. (2006) found transitional housing programs had more selective admissions policies and consequently may serve clients evidencing higher levels of functioning. As a condition of continued stay, transitional housing residents are commonly required to participate in the program's services and training opportunities.

Permanent supportive housing

HUD-funded permanent supportive housing (PSH) is intended to provide housing and appropriate support services for homeless individuals with disabilities including mental illness, physical disabilities substance abuse problems, or AIDS and associated illnesses (HUD, 2002). Wong et al. (2006) point out that “support” is the critical distinguishing feature of PSH as it enables individuals with significant disabling conditions to acquire and retain housing. Martinez and Burt (2006) report that for a sample of 236 single adults, PSH residents had significantly fewer emergency room visits, lower per-person rates of emergency room visits, reduced rates of psychiatric hospitalization, and fewer psychiatric hospitalizations per person.

Wong et al. (2006) found that whereas up to 30% of emergency shelters and transitional housing programs serve families, 61% of permanent supportive housing programs serve only single adults. Further, these authors found that permanent supportive housing programs typically provide more privacy to clients, were more accepting of behavioral problems, and were more lenient in regards to length of stay and program requirements.

One emerging manifestation of permanent supportive housing is the Housing First model. Housing First is an evidence-based approach to providing permanent, independent housing to homeless individuals with mental illness, co-occurring substance abuse, or other health problems by removing barriers to housing entry. For instance, Housing First programs typically do not require sobriety and treatment compliance as prerequisites to housing placement (Stefancic & Tsemberis, 2007). Promotion of consumer choice, recovery, and community integration are goals of Housing First programs. Tsemberis, Gulcur, and Nakae (2004) conducted a study of 225 homeless individuals with a history of mental illness and substance abuse who were randomly assigned to either a control or experimental group. For the control group, the provision of housing was dependent on achieving sobriety and participating in treatment, whereas the experiment group was immediately provided housing without expectation of sobriety or treatment participation. The study participants were interviewed every 6 months over the course of 24 months. Participants in the experimental group had a significantly higher rate of rate of obtaining housing and remaining housed compared to the control group. There was no significant difference in alcohol and drug use between the two groups despite the fact that the control group had a significantly higher use of substance abuse treatment programs. Further, there was no significant difference in psychiatric symptoms between the two groups.

Permanent housing

Permanent housing refers to a house or apartment owned or rented by an individual or family or some other non-transitory living arrangement. For instance, an adult living in the home of his or her parents, without the expectation of finding another housing arrangement, might be said to have permanent housing. For our purposes, permanent housing is distinguished from permanent supportive housing by the absence of formal supportive services provided to the occupant to assist him or her in retaining permanent housing.

Not all homeless individuals need or require supportive services to leave homelessness and remain housed. This may be particularly true for individuals and families with longer histories of stable housing, who then lose their housing owing to economic conditions and transient disruptions such as domestic violence. For these individuals and families, especially those who are first-time homeless, remediation conditions and risk factors that produced the episode of homelessness can be sufficient to facilitate resumption of permanent housing (Shin et al., 1998; Wood, Valdez, Hayashi, & Shen, 1990).

CONCLUSION

In the United States, the efforts to prevent, remediate, and end homelessness extend from local governmental and faith-based endeavors offering food, clothing, employment, and housing to state and federal initiatives to coordinate care and provide suitable housing. The present-day national economic crisis is constraining the financial and material resources available to respond to homelessness while simultaneously increasing the number of individuals at risk of becoming homeless. Efficacious responses to homelessness require an appreciation of the complexity of homelessness; its multiple, interacting causes; diverse manifestations; variable duration; and costly financial and social consequences.

The ecological model of homelessness presented here attempts to represent the domains and complexity of this tragically persistent social phenomenon. Our ecological model, represented in Figure 1, is a conceptual map depicting the relationships and interactions between the constituent elements of the model. Scientist and philosopher Alfred Korzybski famously said, “The map is not the territory.” This insight is especially germane to the map presented in Figure 1. No set of lines can portray the territory of the experienced realities of homelessness, its precursors, and its consequences. Instead, this ecological model of homelessness and its explication in this article is our effort to distill a large body of social science literature into a coherent and cogent map to guide individuals, agencies, and communities in their efforts to prevent and redress homelessness.