Using pension payments abroad to examine post-retirement migration and health among Finnish migrants in Sweden

This prospective cohort study used linked Swedish administrative registers and pension payment data to examine post-retirement migration patterns and health selection among Finnish migrants compared to native-born Swedes. The cohort included individuals born between 1938 and 1955 who were resident in Sweden at age 64, with follow-up from age 65 until death or the end of 2021 (n = 1794,196; 80,313 Finnish-born). Health status was assessed using the Charlson Comorbidity Index (CCI), while pension payments were used to identify country of residence and mortality occurring after emigration. Post-retirement emigration was substantially more common among Finnish migrants (3.0 %) than native-born Swedes (0.3 %), primarily due to return migration to Finland. Contrary to the "salmon bias" hypothesis, poorer health reduced rather than increased the likelihood of emigration for both groups. Among Finnish migrants, lower CCI scores, male sex, being unmarried, and lower income predicted a higher likelihood of post-retirement migration. Sequence analyses revealed three distinct Finnish migration trajectories: early return, late return, and return with onward migration. Mortality did not differ significantly between Finnish migrants who remained in Sweden and those who resided in Finland or other countries; adjusted models even indicated slightly lower mortality among return migrants Extending follow-up abroad with pension data did not change mortality estimates, indicating unrecorded deaths abroad likely did not bias results. The findings indicate positive health selection into return migration and no evidence of a salmon bias effect. Post-retirement migration among Finnish migrants in Sweden is rare and has little impact on migrant-native mortality differences.