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 migrants involves challenges related to their life courses being stretched over several geographies while most data are collected for one specific geography only. Sometimes this produces various “paradoxes” in migrant demography. For example, mortality research high-lights the surprisingly low mortality of migrants to Europe and the US, despite their imminent socioeconomic disadvantages. Three main mechanisms may contribute to these patterns: 1) the so called “salmon effects” in mortality, being produced by situations when migrants opt to return to their country of origin in anticipation of death, 2) selective immigration of healthy migrants, and 3) the underreporting of emigration and return migration. In our study, we test these hypotheses by focusing on the migration effects that may occur within a single country, for which register data can cover the entire life courses of individuals. Instead of focusing on international migrants, we exploit Swedish register data to detect possible salmon and healthy-migrant effects for long-distance migrants from Northern Sweden (Norrland) to Southern Sweden. We apply hazard regression methods to Swedish population registers to examine the age, duration-specific and time-varying influences of migration status on individual mortality during 1971-2007. The study covers 11.9 million Swedish born individuals. About 473,000 people were born in Norrland and had moved to other parts of Sweden, about 131,000 of them had returned to Norrland during our observation period. To distinguish between healthy-migrant and salmon effects we ask whether migrants from Norrland to Southern Sweden have lower mortality than the general population of Norrland, whether they have lower mortality than the population of Southern Sweden, whether return migrants to Norrland have a higher mortality than those who stay in the South, and whether patterns are altered by age. First descriptive results confirm that both mechanisms can be observed for the internal migrants in Sweden.