Herb Westerfelt CONDITIONS CONTRIBUTING TO LONGTERM HOMELESSNESS: An Exploratory Study

In this paper we examine the duration of homeless careers. We build a model of career length based on four conceptual frameworks: institutional disaffiliation, psychological dysfunction, human capital deficit, and cultural identification. Using survey data from a sample of 331 individuals in Minneapolis, we estimate a structural equation model of homeless career onset and duration. We find that conditioned on age, people who have less consistent work histories, experienced childhood foster care, and currently express less discomfort with life on the streets have longer homeless careers. Contrary to our hypothesis we find that people who experienced pre-homeless psychiatric hospitalization had relatively shorter homeless careers and people who suffered from severe symptoms of alcoholism had homeless careers no different in average length than those of other sample members. The Duration of Homeless Careers: An Exploratory Study1 Authors Affiliation During the past decade, homelessness in the United States has become recognized by policy makers, social scientists, and the general citizenry as one of the nation's major social problems. Despite this recognition, the scholarly literature on homelessness remains limited both in quantity and quality. The literature consists almost entirely of univariate descriptions of samples of homeless persons in various communities, case studies of homeless individuals or families, and speculative essays on how changes in employment opportunities, housing markets, and mental health services have altered the size and character of the homeless population. Almost totally absent are systematic empirical studies of these as well as other issues concerning the individual characteristics and experiences that increase the likelihood of becoming homeless, the circumstances leading to lengthy homeless careers, and the conditions enhancing individuals' chances for exiting homelessness. This paper addresses one of the latter concerns, namely the conditions affecting the duration of individuals' homeless careers. Our research is motivated by the hypothesis, advanced but not tested, that individuals with longer homeless careers differ systematically from those with comparatively shorter careers (Redburn and Buss 1986; U.S. Department of Housing and Urban Development 1984).' We focus on the duration of homeless careers rather than the causes of (or vulnerability to) homelessness for three reasons. First, research on vulnerability requires a nonhomeless comparison group, but finding an appropriate contrast group presents extremely complex conceptual and logistical problems whose solutions, if possible, are beyond the scope of this investigation. Second, results of research that predicts duration of homeless careers from characteristics of individuals may have implications for predicting onset of homelessness. That is, if characteristics which affect career duration are also overrepresented among the homeless population in 2 general, we might surmise that they are pertinent to the onset of homelessness as well. Third, the study of homeless careers has substantial policy significance: in providing information on the experiences that prolong homelessness, it may suggest which policies and programs are likely to shorten homeless careers. Our investigation is based on the general assumption that, at the level of the individual, sustained homelessness reflects an estrangement from conventional institutions of society, and that this estrangement comes about in four interrelated ways. First, individuals can suffer a psychological or emotional disorder, which in turn hampers their attempts to participate fully in conventional society (Bogue 1963; Levinson 1957; McCook 1893; Sutherland and Locke 1936). Second, individuals can fail to invest in education, job training, or other forms of human capital, and consequently, remain at a competitive disadvantage in the labor market (Sutherland and Locke 1936; Rooney 1970). Third, individuals can become physically or emotionally disaffiliated from conventional institutions such as the family (Bahr and Caplow 1973). Fourth, individuals can come to identify with street life, becoming "comfortable" in leading a homeless existence rather than a more conventional existence (Wallace 1965). In the analyses to follow, we build and empirically test a model that views homeless career length as a function of the four sources of estrangement outlined above: psychological dysfunction, human capital deficiency, institutional disaffiliation, and cultural identification. We proceed in five sections. After a brief background discussion, we present our theoretical framework and specify testable hypotheses. We follow with a discussion of our sample, variables, and estimation procedures. We then present findings from our analyses and conclude with a discussion of the implications of our results.