Association for Public Policy Analysis & Management

Rural-urban migration is a key component of economic development, and is currently causing concern among policymakers and urban planners in rapidly developing countries such as China and India. While the economic consequences of such migration have been extensively studied, little is known about the health effects of moving to urban areas, especially over the long term.  We estimate the causal effect of rural-urban migration in the United States, focusing on individuals born in the Dakotas and Montana over the period 1916-1927.  As a consequence of the mechanization of agriculture, over half of this group migrated out of this area at young ages, almost all of them to urban areas.  Using a unique source of data containing exact place of birth and death information, we show that despite the positive selection of migrants on characteristics like education, migrants experience decreased longevity in older age compared to non-migrants.  Our results show a reduction in the ten-year survival rate between ages 65 and 75 of 11% for migrants relative to non-migrants.  We control for selection using the amount of mail carried on railroads in 1924 as an instrument for migration.  This instrument is strongly correlated with migration to urban areas, and we believe it to be unlikely correlated with older age mortality. We show that if it is correlated with longevity, the correlation is likely positive.  In this case, our results underestimate the negative effect of migration on longevity, and the reduction in life expectancy for migrants relative to non-migrants is even larger than our IV estimates.  We also estimate marginal treatment effects, which show that the negative effects of rural-urban migration on longevity are largest for lower ability migrants, and higher ability migrants may even gain life expectancy from moving to urban areas.  Our results have important implications for planners and policymakers in countries currently undergoing the urbanization process.