Upgraded to Bad Jobs: Low-Wage Black Women's Relative Status since 1970

Abstract

Labor market changes complicate the analysis of black women's status relative to white women because education, occupational attainment, and race–gender are now less predictive of earnings. Low-wage black women's relative status has improved somewhat from 1970 to 2000, contrary to the well-documented decrease in relative status reported for all black women wage earners since 1980, but their dramatic occupational upgrading was not responsible for the trend. White-collar occupational positions formerly responsible for white women's relative earnings advantage no longer deliver that reward, as restructuring has produced a proliferation of bad jobs across occupational groups. This study argues that increasing exposure to precarious work is crucial to understanding changes in low-wage black women's relative economic status since 1970.

ACKNOWLEDGMENT

We thank Joya Misra, Don Tomaskovic-Devey, Melissa Wooten, and Sal Saporito for their comments on earlier drafts of the article.

NOTES

Notes

¹ Metropolitan areas are good proxies for local labor markets because they are, by definition, functionally interrelated economic units whose limits are defined by commuter flows and other measures of economic interrelatedness. As such, metropolitan areas are the level of aggregation at which local forces of supply and demand operate, and are thus an appropriate unit of analysis for examining earnings from an institutional perspective. See McCall (2001) for a good discussion of local labor market measurement debates.

² This minimum number of individual observations per metropolitan area was calculated by defining an acceptable margin of error of 3 percent and calculating the number of observations necessary in each MSA to meet that standard based on a 95 percent two-tailed confidence interval.

³ We use the terms earnings and wages interchangeably to refer to individuals' hourly wage and salary income from their primary job. Census respondents are not asked to estimate their hourly earnings, so we perform a series of calculations and adjustments to estimate compensation for an hour of each respondent's time. Respondents estimate their total wage and/or salary income for the previous year, as well as the number of weeks worked and usual number of hours worked per week. These variables were used to estimate hourly earnings, which are adjusted for inflation using the CPI-U-RS $2002. Outliers were trimmed, following the procedures described in Mishel et al. (1999).