"A series of proportional hazards models are used to study the relationship between migration history and migration behavior for a sample of young adults from the [U.S.] National Longitudinal Survey of Youth. The results support the argument that migration is a selective process. College educated young adults have a greater hazard rate of making an initial migration but a lower hazard rate of re-migration, suggesting they have less need of corrective geographic behavior. Individuals who have moved two or more times are less responsive to national unemployment conditions than first time migrants. Migration is related to the timing of unemployment within a sojourn. The findings suggest that migrant stock is an important determinant of how labor markets function."