I estimate the impact of Universal E-Verify laws on the college attendance of undocumented Hispanics in the United States. To do so, I implement a series of event studies that account for staggered adoption over time, and I use a random forest algorithm as my primary approach for predicting undocumented status. My results indicate that Universal E-Verify laws lower the college attendance of undocumented Hispanics ages 18-24 by about 3.7 percentage points. This is a substantial effect: only 15.7 percent of undocumented Hispanics ages 18-24 in treated states were enrolled in college following the passage of the laws. This effect is robust to using logical imputation on non-citizen Hispanics to proxy for undocumented immigrants, using a logit model instead of random forest, testing for migration spillover effects on bordering states, and considering potentially confounding impacts of other state-level policies. I develop a theoretical model that explains the mechanisms through which Universal E-Verify affects college education, and I test this model’s implications. I find suggestive evidence that the effect is driven by a negative labor market shock on undocumented adults ages 25-54, which likely leads to worse schooling for their children and renders college less attainable. These findings indicate that employment restrictions targeting working-age undocumented adults hinder the human capital development of undocumented youth.
I explore possible explanations for the residual wage gap between refugees and natives in Germany. I first demonstrate that most of the increased wage gap compared to previous cohorts is not primarily attributable to increased income inequality: a Juhn, Murphy and Pierce (1993) decomposition reveals that residual price changes explain only a small portion of the increased residual gap. I then estimate the impact of prejudice and the minimum wage. While I find limited evidence that prejudice increases the residual wage gap, my results indicate that average prejudice and prejudice at the upper tail of the distribution negatively impacts refugees. This finding diverges from Becker’s model on the “marginal” prejudiced employer, but it is consistent with alternative theories, such as prejudice by customers, or lower reservation wages due to increased job search costs in a prejudiced environment. I also find evidence that increases in the minimum wage bite across regions within Germany lower the residual wage gap without affecting the residual employment gap. Finally, I find that refugees in more prejudiced regions invest more in integration courses and language acquisition. However, refugees in these regions do not earn higher wages, suggesting that prejudice may mitigate potential returns to assimilation.
I examine the impact of the German for Professional Purposes course on the labor market outcomes and human capital of recent refugees in Germany. I focus on language acquisition, employment, skill level, and wages as my key outcomes. I further investigate how the returns to the course are impacted by anti-refugee prejudice. I also test for spillover effects within refugee households. Estimates indicate that this course has remained an effective tool in improving refugee language acquisition. This finding is in the wake of a 2015 crisis that led to a sudden, unexpected, and large influx of asylum seekers, increased prejudice against refugees, settlement policy changes, and a cohort for which the course was not specifically designed. I find evidence that increased prejudice may lower the return to the course. I fail to find evidence for within-household spillovers. My results imply that professional training and language courses can remain an effective policy tool, even in the event of large unexpected crises and upheaval.
Ongoing project with Francine Blau, Lawrence Kahn, Leonardo Penaloza Pacheco, Freddy Bachmann, and Helen Burkhardt.