Migration and its e ects on both migrants and non-migrants has been a focus of social, demographic, and economic research in both developed and developing countries. A major problem faced by all researchers studying the effects of migration is the tendency of migrants to differ from non-migrants on many important characteristics. This paper employs an instrumental variables strategy and a unique source of data to estimate the causal impact of long-distance migration on mortality over age 65. To do so, we consider individuals born in the states of North Dakota, South Dakota, and Montana over the period 1916-1927. This group migrated out of these three rural states at a very high rate, and the vast majority migrated out of the area before the age of 40. We show that migrants have systematically higher education and earnings than non-migrants, two characteristics shown in the literature to increase longevity. To control for the selection on these characteristics as well as other, unobserved ones, we instrument for migration using distance of an individual's place of birth from a railroad line. Our results show that given that one has reached age 65, migrating out of these three states reduces the probability of living to age 75 by 16% compared to those who remain in their area of origin. This finding directly conflicts with the "healthy migrant" hypothesis proposed in the literature, and has implications for countries currently experiencing high internal migration as well as for research investigating the causal relationship between education and health.
|Original language||English (US)|
|State||Published - 2012|