Internal migration, health selection, and the salmon bias: A register-based study of Finland.

Studies on international migrants have repeatedly found a mortality advantage of migrant over native-born populations. Data artifacts, differential prevalence of health-related behaviors, and health-related selection of immigrants and return migrants have been proposed as explanations. Neither the existence of a migrant mortality advantage for internal migrants nor the validity of existing explanations for this group have been extensively studied. Taking advantage of Finnish register data, we extend the literature on health and internal migration in four ways: 1) by using finer geographic units than previous studies, 2) by adopting models that provide more flexibility compared to alternatives based on the proportional hazard assumption, 3) by distinguishing migrants based on whether they return to their birth region (returnees) or do not (leavers), and by age at migration, and 4) by examining cause-specific mortality. We find that both leavers and returnees enjoy a mortality advantage over non-migrants. For both groups, the mortality advantage relative to non-migrants declines with age but is more pronounced for those who move above age 60 and small or negative for those who move at prime working ages. Circulatory-disease mortality accounts for more than half of the longevity advantage of both leavers and returnees. External and alcohol-related causes also contribute, particularly at younger ages. Our results challenge the idea that findings from studies of international migrants can be fully generalized to internal migrants. We demonstrate a consistent healthy migrant effect for all internal migrants, both those who leave and those who return to their region of birth.