Long-Distance Migration and Mortality in Sweden: Testing the Salmon Bias and Healthy Migrant Hypotheses

Sven Drefahl, Stockholm University
Gunnar Andersson, Stockholm University

The demographic study of migrant populations sometimes produces different “paradoxes” as observed demographic rates fail to conform to expectations. The unexpectedly low mortality of immigrant populations in Europe and the U.S. is one such paradox. This is often ascribed to healthy-migrant effects and to “salmon effects” caused by selective return migration. Instead of focusing on international migrants, we test these hypotheses by studying the mortality of long-distance migrants in Sweden, avoiding the problem that individuals disappear from our data at migration. We exploit Swedish register data to distinguish between the two effects in migrant mortality. We cover 11.9 million Swedish born individuals, of whom 473,000 were born in Norrland and moved to other parts of Sweden. 131,000 of them returned to Norrland. Separate analyses are conducted for adult and old age mortality by sex. First descriptive results confirm that both mechanisms can be observed for long-distance migrants in Sweden.