The Return to Caring Skills: Gender, Class, and Occupational Wages in the US

ABSTRACT

Feminist economics has contributed to the understanding of the economic importance of care work. Most studies find a wage penalty associated with caring occupations. This study extends the feminist research on care work beyond caring occupations by identifying specific caring skills and activities derived from the 2014 O*NET job-evaluation data. Four caring skills – (1) Assisting and Caring for Others, (2) Establishing and Maintaining Interpersonal Relationships, (3) Service Orientation, and (4) Social Perceptiveness – were used in ordinary least-squares and quantile wage regressions for 623 occupations in the United States. Findings indicate that the return to caring and assisting skills results in a wage penalty for low-wage workers but a wage premium for workers in high-wage and male occupations. By identifying the impact of gender and class on the economic return to particular caring skills, the study broadens the understanding of care work, especially in relation to US wage inequality.

KEYWORDS:

• Care
• earnings differentials
• gender wage inequality
• labor market inequality
• skills

JEL Codes:

• B54
• J31

ACKNOWLEDGMENTS

The author wishes to acknowledge the anonymous reviewers and Associate Editor of Feminist Economics for their constructive comments and suggestions.

Notes

1 Here, class refers to socioeconomic status rather than the (non)ownership of productive capital assets. It is also used in reference to locations along the wage quantile.

2 O*NET Resource Center, version 19.0 (2014).

3 Skills rated as “not important” were assigned a skill level rating of 0.

4 Because the annual median wage by occupation has a lognormal distribution, the log of wages was calculated and used as the dependent variable. In addition, the OES annual wage is calculated on the basis of full-time, year-round work.

5 Post-regression diagnostics on the variance inflation factor (VIF) resulted in a range from 1.7 to 4.29 with a mean VIF of 2.56, well below the traditional threshold indicators of collinearity. The tolerance indicator for each variable was also well above the 0.1 level, values below are indicative of the presence of multicollinearity (O’Brien 2007). Furthermore, the mean VIF was much higher (8.10) and tolerance indicators near or below 0.1 when separate skill importance and skill-level scores were included in the regression.

6 This designation of female and male occupations is also used in research on occupational segregation (Blau, Brummund, and Liu 2013) and gender inequality (Roth 2004).

7 Summary statistics are shown in Supplemental Table 1, available in the supplemental online appendix.

Additional information

Notes on contributors

Bruce Pietrykowski

Bruce Pietrykowski is Professor of Economics at the University of Michigan–Dearborn and Research Scientist at the Institute for Research on Labor, Employment, and the Economy at the University of Michigan-Ann Arbor. He holds a PhD from the New School for Social Research. He currently directs the Center for Labor and Community Studies at the University of Michigan-Dearborn. His most recent research project focuses on the skills of low-wage workers in the United States.