Abstract
The first chapter presents a series of novel facts about US internal migration based on gravity estimations using detailed county migration data from the Internal Revenue Service. To account for the existence of zero flows in the data, I perform Poisson Pseudo Maximum Likelihood and Tobit estimations in addition to Ordinary Least Squares. I find that migrant flows and incomes are positively related to total incomes in the origin and the destination, and negatively related to the distance between those. Households move from areas with lower average incomes, higher housing costs and unemployment rates to areas that are the opposite. The pull effect of average income on migration is stronger than the push effect, while the push effects of housing cost and unemployment rate are stronger than the pull effects. Additionally, average migrant income increases with average incomes in both the origin and the destination, as well as the distance traveled, and decreases with the share of migrants out of an area. These empirical patterns suggest positive selection of migrants.
The second chapter develops a spatial equilibrium model that accounts for the observed migration patterns, and describes the channels through which import competition affects migration. I employ two-stage least squares estimation to identify the causal effects of rising import exposure on migration, exploiting exogenous variations in lowering trade costs and rising Chinese productivities. I find that a $1,000 increase in imports per worker lowers the population by 5.7 percent, while raising the out-migration rate by 1.7 percentage points in a local labor market. The same increase in imports per worker in the origin raises net migrant outflows by 0.574 log points, and that in the destination lowers net inflows by 0.418 log points. The stronger effect of import exposure in the origin than in the destination implies possible information asymmetry faced by potential migrants who are more familiar with their current residences than future ones.
The third chapter examines the productivity and substitution effects of immigration and offshoring in service sectors in the US. I explore the skill distributions of natives, immigrants and offshore workers to gauge their substitutability. In order to jointly analyze the effects of immigration and offshoring, I match data from the IPUMS samples of the Census-American Community Survey with the multinationals dataset and service imports data from the Bureau of Economic Analysis. I find evidence of substitution between native and offshore workers, and a negative productivity effect of offshoring. Immigration generates positive productivity effect, and immigrants complement native employment. Additionally, the skill complexities of tasks performed by native and immigrant workers increase with more offshoring, and the gap between native and immigrant complexity narrows.
