Employment relationships over time: retention and promotion

In this dissertation, I examine how available information affects promotion and turnover decisions in internal labor markets. In the first essay, I use data on all Major League Baseball managers from 1950 to 1996 to consider multiple evaluation measures and their role in actual firings of managers. The results indicate that firms use all of the distinct measures of managerial performance in termination and rehire decisions. However, the results also suggest that teams, in making termination decisions, use information that is unlikely to reflect managerial ability. That is, talent at the time of hire affects the risk of termination, even after conditioning on team performance relative to expectations after the date of hire. In the second essay, I use the 1979 National Longitudinal Survey of Youth data to explore the factors that are important determinants of an individual’s promotion. One issue that arises in estimating the probability of promotion from longitudinal work history data is that researchers only observe promotion for individuals who remain at a job between interviews. I improve upon earlier studies by using a bivariate probit analysis to correct the bias from partial observability and provide more informative estimates of the promotion process. These new estimates allow differences in promotion rates across demographic groups to be decomposed into differences in the probability of promotion conditional on staying and differences in the probability of staying. In the third essay, we explore the differential patterns of job attachment between men and women by examining how men and women respond to promotion expectations. Using the 1979 National Longitudinal Survey of Youth, we find that early in their career women with low promotion expectations are more likely to stay on a job than corresponding men. We also find that this difference diminishes with experience.