Migrants’ under-Mortality: Using Causes of Death to Disentangle Healthy Migrant Effect and Salmon Bias in Early Twentieth Century Madrid

Diego Ramiro Fariñas, Spanish National Research Council (CSIC)
Michel Oris, Consejo Superior de Investigaciones Científicas (csic)
Stanislao Mazzoni, Consejo Superior de Investigaciones Científicas (CSIC)
Aitor Garcia Aguirre, Spanish National Research Council
Mélanie Bourguignon, UCLouvain

A number of studies in historical demography have highlighted a mortality advantage among migrants in rapidly expanding cities in the 19th and early 20th centuries. This healthy migrant effect is a surprising phenomenon, as it compensates, to some extent, for the disadvantages experienced by most new arrivals. To better understand this phenomenon, we looked at a large city, Madrid between 1905 and 1908. Our database contains nominal information extracted from the 1905 census on 112,296 inhabitants of two districts, one wealthy district, one poor, supplemented by a mortality follow-up until 31 December 1908. In doing so, we collected 5,234 death certificates, with the cause of death systematically recorded. In debates on the causes of under-mortality among migrants, one interpretation emphasises selection at the point of origin, with the healthiest leaving. Another interpretation is that this result is an artefact, as migrants fell ill in unsanitary cities and then returned home, where their deaths were recorded. This is known as the salmon bias. To disentangle these two causalities, we propose an original approach based on the use of causes of death, while controlling for length of residence in the city and various social and environmental characteristics. Indeed, returning home was possible in the case of latent (though fatal) diseases such as tuberculosis, but not in the case of sudden death (such as cerebral haemorrhage, cardiac arrest, heart failure, etc.). Untangling the mortality advantage of migrants based on causes of death has not been done in historical demography and offers new perspectives.

No extended abstract or paper available

 Presented in Session 30. Mortality Inequalities Across Class, Place, and Borders