Blue Collar Jobs versus Pink Collar Jobs: An Analysis of Gender Differences in Occupational Choice Open Access
Downloadable ContentDownload PDF
The degree of gender differences in labor market outcomes, particularly related to occupational choice, is a longstanding theme in the field of labor economics. This dissertation explores several manifestations of those male/female differences, including participation and compensation, using both present-day and historical data for the United States. The first chapter analyzes the relationship between occupational characteristics and the gender pay gap, hypothesizing that occupational choice (as measured through the characteristics of occupations) can explain a significant portion of the difference in average annual earnings between males and females. This study combines several economic theories, including attribute theory and the theory of compensating differentials, into a framework to account for gender differences in levels of occupational characteristics and their returns to earnings. This builds on the premise that male and females have, on average, different tastes and, therefore, different willingness to pay for various occupational attributes. If women exhibit a stronger preference for positive attributes or a stronger distaste for negative attributes, the market returns attached to those attributes could result in a gender earnings differential. Chapter 1 uses the 2016 American Community Survey individual-level data paired with occupational characteristics from the Department of Labor Occupational Information Network (O*NET) database. Seven occupational characteristics are used in this analysis: Brawn, Flexibility, Interaction, Leader, Mental, Pressure, and Social. Multiple regression analyses identify which characteristics are relatively more female or male, in addition to the market valuation (i.e., the compensating differential) of the occupational characteristics based on their returns to earnings. An Oaxaca-Blinder style decomposition including occupational characteristics is used to explore how much of the gender earnings gap can be attributed to gender differences in occupational choice as measured by job attributes. The results demonstrate that there is a significant gender difference in the distribution of several of the above characteristics, particularly Brawn, Social, and Interaction. Furthermore, approximately 13 percent of the average annual logarithmic gender earnings gap can be accounted for by gender differences in the seven occupational characteristics. The second chapter analyzes the trends in labor market outcomes for females in management as compared to females in all occupations. Specifically, it discusses female participation, average annual earnings, the observed female-to-male average logarithmic earnings ratio, and the adjusted female-to-male average logarithmic earnings ratio. The last measure is computed using an Oaxaca-Blinder style decomposition of log annual earnings. Each of these measures is calculated by year from 2000-2016 using the American Community Survey (ACS) data. Finally, three potential explanations for the increase in female relative earnings over the time period under study are explored. The results show that women (both those in management and in other occupations) made gains in their rate of participation and relative earnings over time. While the increase in female participation was greater among women in management occupations, the increase in relative earnings was greater among women in all occupations, particularly after controlling for measurable characteristics. This increase in the female-to-male earnings ratio can be primarily attributed to improvements in female relative earnings within the 27 occupations that comprise the management category, as opposed to changes in female representation across occupations or changes in the relative sizes of occupations within management. The third chapter, which was co-authored with Barry R. Chiswick, is a historical study of free female labor force participation in the pre-Civil War United States. Rates of labor force participation in the US in the second half of the nineteenth century among free women are exceedingly (and implausibly) low, ranging from 10 to 16 percent. This is due, in part, to social perceptions of working women, cultural and societal expectations of females’ role, and lack of accurate or thorough enumeration by Census officials. Chapter 3 presents an augmented free female labor force participation rate for 1860. It is calculated by identifying free women (age 16 and older) who were likely providing informal and unenumerated labor for market production in a family business, that is, unpaid family workers. These individuals are identified as not having a reported occupation, but are likely to be working on the basis of the self-employment occupation of other relatives in their households. Family workers are classified into three categories: farm, merchant, and craft. The inclusion of this class of workers more than triples the free female labor force participation rate, to 56 percent.Additionally, this chapter analyzes the occupational distribution of free females in 1860. Finally, it uses logit regression analysis to determine which characteristics are associated with the likelihood of free women providing formal or informal labor. This includes an analysis of the effects of the prevalence of slavery on free female labor force participation. The greater the extent of slavery in the woman’s county, the less likely she was to have a reported occupation, especially as a housekeeper.