UC Berkeley

Costa Rica had two major demographic changes in the last 20 years. First, Costa Ricans have fewer children. Second, many Nicaraguans have migrated into Costa Rica. Estimating the effects of migration and fertility changes is diffcult. Both migration and fertility decisions are endogenous to labor market and social conditions. To address this concern, in this thesis I use two natural experiments. First, I use a regression discontinuity approach that takes advantage of the exact timing of a 2001 change in child support legislation. I find that birthrates fell signicantly after the change. Single mothers now find it easier to receive child support payments. This affects marital status decisions. Indeed, births within marriage also fell, suggesting a decline in shotgun marriage. Maximilian Kasy and I analyze the impact of changing family structures on the female income distribution in Costa Rica between 1993 and 2009, using decomposition methods based on reweighting, as in [DiNardo et al.,

1996], and infuence function regression, as in [Firpo et al., 2009]. We find that family structures changes before the law change had a general negative effect on incomes. After the law passed, the trend reverses for higher income women, while lower incomes continue to be negatively affected by changing family structures. The second natural experiment is Hurricane Mitch in 1998. The damage it caused to the Nicaraguan countryside generated a surge in Nicaraguan immigrants. This surge located disproportionately in the regions and economic sectors that already had large Nicaraguan participation. I use difference-in-differences to find that Nicaraguan immigration after Mitch had no impact on native labor market outcomes, except for a small negative income eect on low-education men in the regions that share a border with Nicaragua.