In this paper, I study the determinants of the choice to identify as Hispanic among those who could—those whose parents, grandparents, or selves were born in a Spanish-speaking country. I find that individuals with Hispanic ancestry are significantly less likely to self-identify as Hispanic if they live in states with high levels of implicit ethnic bias. A one standard deviation increase in bias decreases self-reported Hispanic identity by seven and 13 percentage points for first and second-generation Hispanics, respectively. These effects are more prominent among second-generation immigrants whose mothers and fathers were born in a Spanish-speaking country than among children of inter-ethnic parents.

In this paper, I compare the children of inter-ethnic marriages to study the impact of having a Hispanic last name. While males born to Hispanic father-White mothers earn less than those born to White father-Hispanic mothers, the gap could be completely explained by educational differences. I also study the effect of identifying as Hispanic on earnings. I find that men that identify as Hispanic earn significantly less than those that do not, even after controlling for educational differences.

What is the effect of democracy on ethnic favoritism? I estimate the relationship between co-ethnicity and five outcomes of public good provision – education, infant health, wealth, access to clean drinking water and access to electricity – using data from twenty-one African countries. Following previous research, I use variation in co-ethnicity across ethnic groups and over time. I first estimate the relationship between these five outcomes and co-ethnicity in the full sample and then split the analysis between anocracies and democracies. I find mixed evidence of ethnic favoritism in the full sample. I find more evidence of co-ethnic targeting when comparing within democracies only and within anocracies only.