نقش تخصیص کار و موقوفات سرمایه انسانی در توضیح تفاوت های جنسیتی در عملکرد و ارتقای شغلی

We test a job ladders theory of career progression within internal labor markets as developed by Lazear and Rosen (1990). The theory argues that gender promotion gaps are due to sorting of men and women into career tracks with different promotion opportunities based on ex ante quit probabilities. Analyzing US federal government employees using a dynamic unobserved panel data model, we find that job assignment is one of the strongest predictors of gender differences in promotion. We also find that women have to jump higher performance hurdles to promote across grades, but, within grades, their promotion probabilities are comparable to those of men. In this organization, women can be found in both fast- and slow-track jobs, based on their promotion history, suggesting that unobserved heterogeneity is revealed to the firm over the worker's career.

Research on the gender pay gap has recently focused on career differences within large hierarchical organizations. In part this trend has been driven by limited information on human capital endowments in standard public-use data (Donald and Hamermesh, 2004).1 The data limitations present obstacles to distinguishing among competing theories on gender pay gaps, including discrimination, occupational sorting, and job assignment. Separating unobservable characteristics, such as ability or quit propensities, from discrimination and endowment differences has been especially difficult. While women's labor force participation, educational attainment, and representation in traditionally-male occupations has risen dramatically in recent years, the question remains: will these changes be enough to eradicate gender-based differences in wages and career progression? Following the recent trend of studying male–female career differences in the context of firm-level decisions on optimal incentive pay and promotions, we focus on professional workers in a large hierarchical organization. Using longitudinal data on U.S. federal government employees, we test hypotheses generated from the Lazear and Rosen (1990) jobs-based model of gender differences in career progression. The model suggests that gender pay differences arise from the assignment of males to ‘fast track’ jobs that require heavier investments in specific human capital and thus stronger job attachment. Because women have higher productivity in non-market activities, they are more likely to separate and, therefore, are assigned to jobs that have flatter career paths. Using a dynamic unobserved panel data model, we find that job assignment is one of the strongest predictors of gender differences in promotions. However, we also find that women can be on both fast and slow tracks, based on their promotion history, suggesting that the separation probability is revealed to the firm over the worker's career. Contributing to the internal labor market literature more generally, we find that promotions resulting only in wage growth are characterized by different dynamics than promotions involving a change in responsibilities.

Our results suggest that gender differences in careers are driven by a combination of observed variables, most notably tenure, occupation, and position in the hierarchy, as well as unobserved heterogeneity. We find evidence that: (1) females promote more often because we initially observe them in lower-level positions; (2) females need higher performance ratings to move up the hierarchy; (3) within narrowly-defined jobs females can be both on fast and slow tracks, based on their history of promotions; (4) position in the hierarchy explains the largest portion of the gender gap in promotion probabilities; and (5) that unobservables work against women's promotion probability, consistent with the hypothesis that females have higher productivity in non-market activities. Given that these models are equivalent to wage models (individual pay growth occurs primarily via grade changes), the L&R hypotheses with respect to gender wage differences are largely supported. In investigating ‘glass ceilings’ in the federal service, we find evidence of ‘sticky floors.’26 In our data, females are less likely to be in managerial positions primarily because they are concentrated in lower pay grades. However, within the same pay grade level (and also conditional on performance, specific and general human capital, occupation, and unobserved heterogeneity), women's promotion probabilities are similar to or better than those of men, suggesting that discrimination does not explain why women are concentrated in lower-level positions. An alternative explanation following Winter-Ebmer and Zweimuller is that women are assigned to lower positions upon hiring. We explored this hypothesis by analyzing workers who were hired in 1986. We find that women do enter at lower grades, but endowments explain most of these differences.27 As a result, while human capital endowments are important in explaining gender differences in career progression, job assignment is one of the strongest factors driving promotions to higher grades (pay differences) and to supervisory positions among professional workers. One additional contribution of this paper is that it separates the effects of promotions that result in wage growth from promotions that involve a change in duties and responsibilities. Due to less-clear definitions of hierarchical steps and job titles in most prior studies using internal firm data, the literature often combines both events in the definition of a promotion. However, incentives and specific human capital may have varying importance for each type of promotion. Our data allow us to disentangle these effects and show that title-promotions, which involve a change in responsibilities, are characterized by entirely different dynamics than promotions that involve only salary growth. Most importantly, while administrative constraints (such as ‘green card’ effects, or time-in-grade restrictions) may induce negative correlation in wage growth over time, promotions that involve changes in responsibilities (and require different skills or higher ability) are characterized by substantial positive persistence over time.