Mapping trends: patterns of employment trajectories in the United States, 1967-2005.

Abstract

This thesis systematically examines the overall patterns of employment histories in the United States based on data from the Panel Study of Income Dynamics (1967- 2005). Over 15,000 individual employment trajectories, including those from previously under-studied social groups such as workers in the secondary labor market, blue-collar occupations, and those outside big firms, were compared by optimal matching analysis and cluster analysis. Accounting for gender, age period, and birth cohort, the distance of Americans’ career paths from the stable, full-time trajectories was found to be decreasing over time when careers were examined holistically; meanwhile, career paths have become more homogeneous to one another. Gender gap in employment trajectories has considerably declined. These trends have been driven primarily by women’s increasing engagement in full-time employment while men’s trajectories have changed only slightly. Three primary patterns of employment trajectories were found—an unstable, part-time pattern, a stable, inactive pattern, and a stable, full-time pattern that accounted for about three-quarter of the cases. Those in the stable, full-time pattern, particularly men, were highly likely to continue this type of employment trajectories into older ages. A modest occupation effect was found in which white-collar workers and those in occupations dominated by men, such as managers and administrators and machine operatives, were more positively associated with the stable, full-time pattern than those from other occupations, even after gender was taken into account. Overall, this study draws an optimistic picture of employment trajectories that contradicts current perceptions about the disappearance of long-term, full-time employment. Its conclusions, however, need to be interpreted with caution given a possible under-estimation of job changes, part-time work, and underemployment as well as an under-representation of women, economically less active people, and those in big family units due to data quality issues.