| Abstract: | A longitudinal approach to migration behavior makes it possible to identify a sufficient condition for positive associations between rates of in-migration and rates of out-migration in the same regions. The longitudinal approach centers on the intervals of time that individuals spend in a region, and these intervals can be analyzed in terms of probability distributions or the equivalent hazard functions or survivor functions. Differences in the distributions of these intervals between subpopulations whose residence in a region begins with in-migration and subpopulations whose residence begins with other events are sufficient for temporal variations in rates of in-migration to produce variations in out-migration rates in subsequent periods of time. Tests for such differences are performed using data from the National Longitudinal Survey of Youth. |
