Job Market Paper

Abstract: Deportations affect undocumented migrants already in the US and may also affect future migration; relationship potentially mediated by migratory networks. I study the effects of deportations on current and prospective migrants by combining micro-level data on undocumented migrants from Mexico and nationally representative US survey data. I then use an IV strategy that exploits local variation in deportations driven by the staggered rollout of the Secure Communities program (2008-2013), the size of the local undocumented population before activation, and the extent of local authorities’ engagement with the program. Estimates suggest that deportations reduce local undocumented populations more than one-for-one, partly through internal relocations. Long-term, local immigration enforcement reduces the probability that new immigrants settle in a given location even when there is a network connection, and it reduces the probability that incoming flows form new network links. Together, this establishes migrant network disruptions. Moving on new migration, I use a shift-share IV design and data on millions of new undocumented immigrants to estimate the impact of increased exposure to immigration enforcement at the origin via networks. I find that such exposure reduces migrant outflows, indicating a direct deterrence effect. I also find that exposure increases new migrants’ educational attainment, expected wages upon arrival, and the proportion of incoming migrants who speak English, with no effect on the age and gender mix. These findings suggest that migrant networks are widespread and active in the US, and that local immigration enforcement permanently shifts migratory networks, generating new regional migration patterns.