Theses Doctoral

Essays on Gender Differences in Educational and Labor Market Outcomes

Steingrimsdottir, Herdis

With women's increased education and labor market participation in the last few decades the labor market has changed considerably. At the same time the interaction between household activities and work have been constantly evolving, affecting household dynamics and family outcomes, such as fertility, marriage and divorce. The first chapter explores the effect of unrestricted access to the birth control pill on young people's career plans, using annual surveys of college freshmen from 1968 to 1980. In particular it addresses the question of who was affected by the introduction of the birth control pill by looking at career plans of both men and women, and by separating the effect by level of academic ability and race. The results show that unrestricted access to the pill caused high ability women to move towards occupations with higher wages, higher occupational prestige scores and higher male ratios. The estimated effects for women with low grades and from low selectivity colleges are in the opposite direction. Men were also affected by unrestricted access to the pill, as their aspirations shifted towards traditionally male dominated occupations, across all ability groups. The biggest effect of unrestricted access to the pill is found to be on non-white students, both among men and women. The paper uses Census Data to compare the changes in career plans to actual changes in labor market outcomes. When looking at the actual career outcomes, early access to the pill affects both men and women -- shifting their careers towards traditionally male dominated occupations associated with higher wages. Early access to the pill is also associated with significantly higher actual income for men. In the second chapter I look at the relationship between increased access to reliable fertility controls and men's disappearance from teaching. As the pill has been found to have a substantial effect on women's family responsibilities, career investments and labor market outcomes, men's bargaining position in the marriage market is likely to have changed considerably. Teaching stands out among the career choices of male college freshmen in terms of average income and prestige. The effect of the shift in bargaining power on men's career choices is hence likely to be prominent in the teaching sector. Between 1968 and 1980, the ratio of male college freshmen planning to become a teacher fell from 12.4% to 2.4% and the share of males among those who aspired to teach dropped from 30.6% to 19.7%. Using nationally representative data on the career plans of college freshmen I find that unrestricted access to the birth control pill bears a negative relation to the likelihood that men plan to teach, while changes in the strength of teacher unions and relative wages of teachers have limited effect on their career plans. Men's aspirations shift away from teaching towards occupations that are associated with higher average income like accounting and computer programming. The results are supported by equivalent findings looking at actual career outcomes in the Census Data. The third chapter focuses on the role of discrimination and the possibility that education as a tool to reveal ability is more important among women than men. As social networks tend to run along gender lines and managers in the labor market are predominantly male, it may be more difficult for women to signal their ability without college credentials. Moreover, women may use education to signal their labor market attachment. A game theoretical model of racial discrimination and educational sorting, introduced by Lang and Manove (2011) is applied to examine the gender gap in schooling attainment. As the gender gap differs between demography groups, being more prominent for blacks and Hispanics, the model is estimated separately for each race or ethnicity group. Using data from the NLSY79, the results in the paper are consistent with a model where education is more valuable to women, due to signaling. As predicted by the model, education as a function of ability (measured with AFQT scores) is more concave for women than for men. For over 88 % of the whites in the sample women choose higher level of education given their ability, than do men. On the other hand, the model fits the data better for whites than for blacks and Hispanics, and therefore fails to explain the observed differences across race and ethnicity groups.


  • thumnail for Steingrimsdottir_columbia_0054D_10632.pdf Steingrimsdottir_columbia_0054D_10632.pdf application/pdf 3.72 MB Download File

More About This Work

Academic Units
Thesis Advisors
O'Flaherty, Brendan Andrew
Ph.D., Columbia University
Published Here
April 19, 2012