Using mobility networks to understand occupational sex segregation

Decades after the beginning of the gender revolution, most women and men still work in sex-typed occupations. This is a primary driver of the gender wage gap. Research describing the patterns of occupational sex segregation focuses on supposedly innate job characteristics that match gender stereotypical abilities and preferences, such as the use of mathematical skills or social skills, on income and status differences between occupations, and on organizational job characteristics, for example, the need to work long hours. However, beyond such occupational attributes, sex segregation is hypothesized to exhibit emergent patterns that are linked to the interdependent job mobility of women and men, in particular, men selectively leaving feminizing occupations. Developing new tools inspired by statistical network research, and using representative, longitudinal data that contain detailed occupational mobility from the UK between 2000 and 2008, this replacement mechanism is analyzed. I show that 19-28% of observed sex segregation is linked to this emergent phenomenon in a statistical model that disentangles the various predictors of the allocation of women and men to different occupations. This makes it the most important predictor of segregation in contrast to concurrently modelled explanations based on occupational characteristics.

Marginalized by race and place

Purpose – Given South Africa's apartheid history, studies have primarily focused on racial discrimination in employment outcomes, with lesser attention paid to gender and context. The purpose of this paper is to fill an important gap by examining the combined effect of macro- and micro-level factors on occupational sex segregation in post-apartheid South Africa. Intersections by race are also explored. Design/methodology/approach – A multilevel multinomial logistic regression is used to examine the influence of various supply and demand variables on women's placement in white- and blue-collar male-dominated occupations. Data from the 2001 Census and other published sources are used, with women nested in magisterial districts. Findings – Demand-side results indicate that service sector specialization augments differentiation by increasing women's opportunities in both white-collar male- and female-dominated occupations. Contrary to expectations, urban residence does not influence women's, particularly African women's, placement in any male-type positions, although Whites (white-collar) and Coloureds (blue-collar) fare better. Supply side human capital models are supported in general with African women receiving higher returns from education relative to others, although theories of “maternal incompatibility” are partially disproved. Finally, among all racial groups, African women are least likely to be employed in any male-dominated occupations, highlighting their marginalization and sustained discrimination in the labour market. Practical implications – An analysis of women's placement in white- and blue-collar male-dominated occupations by race provides practical information to design equitable work policies by gender and race. Social implications – Sex-typing of occupations has deleterious consequences such as lower security, wage differentials, and fewer prospects for promotion, that in turn increase labour market rigidity, reduce economic efficiency, and bar women from reaching their full potential. Originality/value – Very few empirical studies have examined occupational sex segregation (using detailed three-digit data) in developing countries, including South Africa. Methodologically, the paper uses multilevel techniques to correctly estimate ways in which context influences individual outcomes. Finally, it contributes to the literature on intersectionality by examining how gender and race sustain systems of inequality.

Occupational sex segregation and working time: Regional evidence from Germany

This paper provides descriptive evidence for declining occupational sex segregation on the German labor market, especially concerning the regional differences between the former East and West Germany. I use segregation measures and long-run social security data for the decade of 1992 to 2004. While segregation has declined over time, it remains higher for the eastern part of Germany. Although this finding is observable for full-time and part-time work, segregation is always lower in part-time employment.