Popular perception holds that employment stability has decreased in recent decades. However, no conclusive evidence exists on secular declines in the length of jobs held. Furthermore, most studies conclude that the proportion of long term jobs has remained remarkably stable over the last few decades. To shed light on this discrepancy we use distribution analysis to systematically track changes in Canadian employment durations over an extended period. This is done in order to reconcile popular perception with recent studies and nest the existing literature in a broader historical context. Using finite mixture decomposition on successive cohorts of workers starting from the 1950s we identify worker types within cohort-based distributions. Then, using tests of stochastic dominance, we show that the distribution of employment has indeed changed. The finite mixture decomposition reveals that earlier cohorts were more likely to have longer tenure than later cohorts and that there are shifts in pro-portions between longer and shorter work episodes. Our results also indicate that after the 1960s employment durations declined sharply for men, while for women the results were mixed.

Additional Metadata
Keywords Nonparametrics, Stochastic Dominance, Kolmogorov-Smirnov type statistic, Bootstrap, Heterogeneous Distribution, Censored Distributions, Finite Mixtures, Employment Duration.
JEL Labor Economics: General (jel J01), Semiparametric and Nonparametric Methods (jel C14), Hypothesis Testing (jel C12), Statistical Distributions (jel C16), Duration Analysis (jel C41)
Publisher Department of Economics
Series Carleton Economic Papers (CEP)
Ignaczak, Luke, & Voia, M.-C. (2009). A Nonparametric Analysis Of Canadian Employment Patterns (No. CEP 09-01). Carleton Economic Papers (CEP). Department of Economics.