Immigration and voting patterns : evidence from the U.S. presidential and congressional elections

In this paper I study the relationship between immigration and voting patterns in U.S. Presidential and Congressional elections. I use county-level data on demographics such as race, income, unemployment, and education combined immigration measures on non-naturalized immigrant stock, non-naturalized immigrant flow, and naturalized immigrant flow. Using OLS regression models, I study the effect of these immigration measures on voting for Presidential elections (2000, 2008, 2016) and Congressional elections (2000, 2010, 2016). Controlling for year and county fixed effects, I find that non-naturalized immigrant stock is a significant and positive predictor for share of votes in U.S. Presidential and Congressional elections, nonnaturalized immigrant flow is a significant and negative predictor only for Senate elections, and naturalized immigrant stock is a significant and negative predictor for both Presidential and House of Representative elections