Determinants of Spells of Poverty Following Divorce

Abstract

Introduction Between 1967 and 1984, the proportion of single-parent families increased from 10 to 21 percent of all families (Congressional Budget Office, 1985). The poverty rate of all female-headed families in 1983 was 28 percent, compared to 8 percent for two-parent families. Female-headed households exhibit higher poverty rates than the elderly and disabled persons as well. Over one-half of black female-headed households fell below the poverty line in the same year (U.S. Bureau of the Census, 1983). Although being a member of a female-headed household is not synonymous with being in poverty, the likelihood of a single-mother family being poor is much greater than for other demographic groups. In addition, the gap in the poverty rates between single-mother households and other groups has been widening over time (McLanahan, 1985; McLanahan, et al., 1987). This feminization of poverty has been the result of two factors: the improving economic health of other poor groups over time and the increasing proportion of people living in single-mother families (McLanahan and Garfinkel, 1989; Fuchs, 1986). While the feminization of poverty is a useful conceptual tool, the issue remains as to its root causes. Changes of family structure are by far the major cause of initial spells of poverty among female-headed households. Kniesner, et al. (1988) found that changes in family structure precede poverty spells in over 99 percent of the white single-mother families they studied, and in almost 97 percent of black families. Bane (1985) and Bane and Ellwood (1986), looking specifically at divorce, have estimated that 20 to 25 percent of the poor in the 1970's became poor as a result of changes in their family structure resulting from divorce. In a study of the economic health of women following divorce, Peterson (1985) found that failure to accumulate human capital was a significant factor in determining poverty status. He also found that, in general, nonwhites were more likely to be poor than whites and that human capital investments had less of an impact on the economic health of black female heads of household than their white counterparts. In addition, the holding of nontraditional attitudes about sex roles was correlated, ceteris paribus, with economic health following divorce. In this paper, we examine length of poverty spells following marital disruption for the resulting single-mother household. While never-married heads of household are also at economic risk, changes in family structure, particularly divorce, as discussed above, are significant determinants of poverty. For this reason, our analysis will focus on divorced women. Particular emphasis is placed on racial and cohort differences and the effect of women's accumulation of human capital on poverty status. Data and Model This analysis of the issues related to poverty following divorce uses data from the National Longitudinal Surveys (NLS). The National Longitudinal Surveys consist of five age-sex subsets of the United States population represented by the samples mature men/women, young men/women, and youth. Each of the five groups is represented by a national probability sample of approximately 5,000 individuals (1,500 nonwhites and 3,500 whites). The Center for Human Resource Research at the Ohio State University generates the data by conducting interviews at least once every two years. The first sample comes from the NLS for Young Women and consists of women who were aged 14 to 28 in 1968. The sample contains women who were married as of 1968 and divorced before 1982, or single (or divorced) in 1968 but (re)married and divorced (again) before 1982. For individuals who were divorced more than once, the latest divorce is used in the analysis. The sample size consists of 192 whites and 99 nonwhites. The second sample is taken from the NLS for Mature Women, women who were 30 to 44 in 1967. The same criteria for inclusion in the sample (vis-a-vis changes in marital status) as outlined above were used. …