| This dissertation examines the effects of prior employment histories on subsequent labor market outcomes. Job satisfaction, tenure, promotion opportunities, earnings and non-pecuniary benefits can all be considered as desirable outcomes of the labor market activities. From this list, labor earnings provide a tangible way to evaluate labor market outcomes through remuneration for the work, and are the focus of this dissertation. Empirical analysis is performed using the National Longitudinal Survey of Youth, a nationally representative sample of men and women, interviewed from 1979 to 2006.;First, I examine the effects of intermittent employment histories on earnings by revisiting the traditional Mincerian earnings equation. Recognizing the limitations of the traditional experience measure introduced by Mincer (1958, 1974), I include additional variables in the earnings function to capture gaps in individuals' employment histories. While this ameliorates the omitted variables problem, it introduces an endogeneity concern, as experience gaps are likely correlated with other regressors and the unobserved error term. I overcome these concerns by implementing an instrumental variables (IV) approach. I also correct for the sample selection as a means to adjust for non-random participation in the labor market, which is especially important for the female subsample. The fit of the model improves when accumulated previous employment gaps and most recent unemployment spells are included into the specification along with the potential experience measure. Empirical results suggest that gaps in the experience belong in the earnings function. However, the use of the traditional potential experience measure alone can be justified in the estimations for men over thirty, since their employment histories are the least interrupted. Moreover, the negative effects of previous years not working are more prominent for younger adults, and females are more heavily penalized for each year not working. Performing Blinder-Oaxaca decomposition, I find that improving the specification of the wage equation to account for the heterogeneous employment histories, quality of the acquired experience and selection into the labor market, decreases the "unexplained" part of the gender wage differential by as much as 50% in some specifications.;Additionally, I analyze the effects of frequent job changes on earnings. I define a job change as associated with a change of an employer (and not necessarily with the promotion within the same company), and the record of all reported jobs/employers, along with the duration of unemployment spells, constitute the employment histories I am considering in my research. To properly examine the effects of the job mobility on earnings, I am augmenting the traditional earnings function with the additional measures of continuity of work experience already introduced in my work. Along with the number of years of labor market attachment and time spent not working, I am using the level of job mobility (based on the number of job holdings), degree of labor force attachment, and an indicator of multiple job holdings. The results suggest that employment histories and degree of labor force attachment are important determinants of labor earnings. After the unobserved individual heterogeneity is purged via fixed effects, I find evidence that there are some benefits associated with moderate job mobility at the early stages of the career for men and women. However, very high mobility is associated with higher earnings of highly educated women and those with established careers. |
