Estimating the Effect of School Quality on Mortality in the Presence of Migration: Evidence from the Jim Crow South

Abstract

We consider the effects of school quality on health using the introduction of the Rosenwald schools which expanded educational opportunities for African Americans living in rural areas of the South during the early part of the 20th century. Prior research has shown that the Great Migration dramatically reduced the life expectancy of blacks who migrated north and prior research has also shown that Rosenwald schooling increased the likelihood of migration out of the South. This clouds measurement of a direct channel between schooling and survival. Because of this, we estimate the effect of the Rosenwald schools on mortality conditional on migration status. We find that addressing the endogenous migration of blacks is empirically important. Even though the advent of Rosenwald schools exogenously increased education, there is no correlation between Rosenwald school exposure and survival. But accounting for the negative impact of migration on survival induced by Rosenwald school exposure, we find that that complete exposure to the Rosenwald schools during childhood increased life expectancy by about two to three months. These findings highlight the importance of interpreting IV results carefully within the instruments historical context when addressing underlying mechanisms.