The Review of Black Political Economy

, Volume 18, Issue 4, pp 55–68 | Cite as

The effect of human capital on the economic status of divorced and separated women: Differences by race

  • Teresa Mauldin
  • Joan Koonce


This study investigated the impact of investments in human capital on the economic well-being of black and white women immediately following marital disruption. It also explored the extent to which the observed differences in income between the two groups were due to differences in the levels of qualities (endowments) or differences in the impact of these qualities (discrimination). The average differences in endowments explained almost two-thirds of the income gap between black and white women. Most of this explanatory power was due to differences in educational attainment, work experience, and region.


Human Capital White Woman Black Woman Capita Income Human Capital Investment 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    U.S. Department of Commerce, Bureau of the Census, Current Population Reports,Marital Status and Living Arrangements: March 1986, Series P-20, #418, (December 1987).Google Scholar
  2. 2.
    See, for example, M.J. Bane and R.S. Weiss, “Alone Together: The Work of Single-Parent Families,”American Demographics, 2(May 1980), pp. 11–14,48; A.J. Cherlin,Marriage Divorce Remarriage (Cambridge, MA: Harvard University Press, 1981); T.J. Espendshade, “The Economic Consequences of Divorce,”Journal of Marriage and the Family, 14(August, 1979), pp. 615-625; and L.J. Weitzman,The Divorce Revolution: The Unexpected Social and Economic Consequences for Women and Children in America (New York: The Free Press, 1985).Google Scholar
  3. 3.
    G.J. Duncan,Years of Poverty, Years of Plenty;The Changing Economic Fortunes of American Workers and Families (Ann Arbor, MI: Institute for Social Research, 1984).Google Scholar
  4. 4.
    T.J. Kniesner, M.B. McElroy and S.P. Wilcox, “Getting into Poverty Without a Husband, and Getting Out, With or Without,”American Economic Review, 78(May, 1988), pp. 86–90.Google Scholar
  5. 5.
    See, for example, M.J. Bane, “Marital Disruption and the Lives of Children,”Journal of Social Issues, 32(1976), pp. 103–117; G.S. Masnick and M.J. Bane,The Nation’s Families: 1960–1990 (Cambridge: The Joint Center for Urban Studies of MIT and Harvard University, 1980); N.M. Rudd and M.M. Sanik, “Inappropriate Work Plans of Young Women: Causes and Consequences,” in K.D. Rettig and M. Abdel-Ghany (eds.),Economic Decisions of Families: Security for the Elderly and Labor Force Par- ticipation for Women (Washington: American Home Economic Association, 1983); and J.W. Smith and E.S. Beninger, “Women’s Nonmarket Labor: Dissolution of Marriage and Opportunity Cost,”Journal of Family Issues, 3(June, 1982), pp. 251-265.CrossRefGoogle Scholar
  6. 6.
    G.J. Duncan and S.D. Hoffman, “A Reconsideration of the Economic Consequences of Marital Dissolution,”Demography, 22(November, 1985), pp. 485–497.CrossRefGoogle Scholar
  7. 7.
    Cherlin, 1981.Google Scholar
  8. 8.
    G. Becker,Human Capital (New York: Columbia University Press, 1964).Google Scholar
  9. 9.
    U.S. Department of Commerce, Bureau of the Census, Current Population Reports,Educational Attainment in the United States: March 1987 and 1986, Series P-20, #428, (August 1988). Ages of women in the sample at time of first divorce or separation ranged from 17 to 36. The mean age at divorce was 26.35 years of age for white women and 25.2 years of age for black women.Google Scholar
  10. 10.
    See, for example, U.S. Department of Commerce, Bureau of the Census, Current Population Reports,Poverty in the United States 1985, Series P-60, #158, (October 1987); R. Farley and W.R. Allen,The Color Line and the Quality of Life in America (New York: Russell Sage Foundation, 1987).Google Scholar
  11. 11.
    It was not possible in this study to separate the effect of human capital on women’s own earnings from its effect on nonearned income (e.g., child support) because detailed data on nonearned income were not recorded during the early years of the NLS interviews. Had such data been available, it would have been appropriate to specify separate models for the two sources of income. The earnings model would then have consisted of two parts: an equation estimating the probability of labor force participation and an earnings equation which controlled for that probability. Since appropriate data were not available, a single equation is specified with total household income (in per capita terms) as the dependent variable, and measures of human capital and control variables (including labor force participation) as the independent variables. Regression results thus reflect the impact of human capital on total income and not just its effect on earnings. We estimated a two-equation model including a labor supply equation and an income equation which included predicted labor supply as an independent variable. In the income equation the coefficient of predicted labor supply was negative, a counterintuitive result which suggests that nonearned income is playing a role in labor supply decisions immediately following marital disruption.Google Scholar
  12. 12.
    One possible measure of weeks worked for this study was the sum of weeks worked from the 1968 survey until disruption. However, this variable was extremely limited by truncated records before 1968 and every-other-year surveys after 1973. For those women who divorced early in the surveys and for those at the oldest end of the cohort who could have worked up to six years before the surveys started, truncated records posed a serious problem. Because of the existence of alternate year surveys after 1973, younger women would show fewer weeks worked than they may have actually worked. Thus, measurement error would be high with this variable. An alternative, predicted weeks worked, was estimated by regressing actual weeks worked in any year on years of education, occupational status, age, race, husband’s income, and number of children for all respondents in that year. A logistic function was utilized so that predictions would be constrained to 52 weeks per year.Google Scholar
  13. 13.
    A.S. Blinder, “ Wage Discrimination: Reduced Form and Structural Estimates,”Journal of Human Resources, 8(Fall 1973), pp. 436–455 and R. Oaxaca, “Male-Female Wage Differentials in Urban Labor Markets,”International Economics Review, 14(October 1973), pp. 693-709.CrossRefGoogle Scholar
  14. 14.
    The poverty threshold statistics published by the Bureau of the Census were utilized to compute the poverty/nonpoverty status of the women. The measure of poverty status used in this study, then, was a ratio of actual income (pre- and post-) to some established minimal income for a certain number of family members.Google Scholar
  15. 15.
    Because of high correlation (r =.96) between occupational status and work experience, occupational status was dropped from the analysis. See R. Farley and W.R. Allen,The Color Line and the Quality of Life in America (New York: Russell Sage Foundation, 1987) for a discussion of occupational differences by race.Google Scholar

Copyright information

© Springer 1990

Authors and Affiliations

  • Teresa Mauldin
  • Joan Koonce

There are no affiliations available

Personalised recommendations