How did divorced women fare in the 1990s?

Abstract

This paper presents evidence of changes in employment and real wages in the population of divorced single women during the 1990s. Using data from the U.S. Current Population Survey (CPS) for 1989 and 1999, the paper estimates multivariate models of labor force participation and hourly wages for each year. Differences between years in employment and wages are decomposed into portions attributable to changes in measured characteristics and changes in coefficients of the models. Estimates indicate that full time employment remained virtually unchanged during the decade, and real wages increased by less than 2%. Decomposition of the regression models shows that measured characteristics in this population changed in a direction that would have lead to higher wage growth, but those changes were offset by changes in the model’s coefficients. The result is that earnings experienced only modest growth. In the labor force participation model, changes in measured characteristics worked in the direction of a modest decrease in full time employment, but again coefficient changes provided an offsetting effect.

Acknowledgements

An earlier version of this paper was presented at the 2002 conference of the Missouri Valley Economic Association. The author thanks Linda Ghent for helpful comments.

Notes

1 The Census Bureau implemented a redesign of the Current Population Survey in 1994. Included in the design were reformulated questions about earnings. Cohany, Polivka, and Rothgeb (1994, Table 7) estimate that the revised questions produce slightly higher estimated earnings. For all full time workers, the difference is approximately 3%. For the subset of full time women, however, they did not find a difference.

2 In 1992 the Census Bureau redesigned the question concerning years of schooling. The revised question elicited categories of educational attainment rather than years of schooling completed. In order to express education as years completed, I used the following conversion for 1999:

3 Due to restrictions of sample size, I consolidated the following industry categories: (1) manufacturing, mining, agriculture and construction were combined into “manufacturing”; (2) entertainment and professional services were combined into “professional and related services.” Similarly, I combined the following occupations into “skilled production”: precision, production, craft and repair occupations; and farming, forestry and fishing. Table A.1 contains precise definitions of the industry, region and occupation dummy variables.