Long-duration employment opportunities are a necessary condition for workers to hold lifetime jobs. This paper uses longitudinal data on individual U.S. manufacturing plants from 1963-1982 to estimate the age and completed spell distributions for employment positions. The results indicate that, of the employment opportunities in progress in the U.S. manufacturing sector in 1982, 30.0% were at least 19 years old and 59.6% would have a completed length of at least 20 years. High rates of turnover in employment positions coexist with a large number of long-duration employment opportunities because the turnover tends to be concentrated within a subset of the producers.

In recent years a growing number of countries have constructed data series on job creation and job destruction using establishment-level data sets. This paper provides a description and detailed comparison of these new data series for the United States and Canada. Pint, the Canadian and United States industry-level job creation and destruction rates are remarkably similar. Industries with high (low) job creation in the U.S. exhibit high (low) job creation in Canada. The same is true for job destruction. In addition, the overall magnitude of gross job flows in the two countries is comparable. Second, the time-series patterns of creation and destruction are qualitatively similar but do differ in a number of important respects. In both countries, job destruction is much more cyclically volatile than job creation. This cyclical asymmetry is, however, more pronounced in the United States. The paper finishes with a characterization of the job flow patterns using a modified