Association for Public Policy Analysis & Management

Matthew Hall, Cornell University

Jacob Rugh, Brigham Young University

In response to the rise in the number of unauthorized immigrants over the last several decades, local communities have taken unprecedented steps to facilitate immigrant detentions and deportations. Since the early 2000s, the number of deportations has soared, rising from 165,168 in 2001 to 438,421 in 2013. These activities have further intensified by recent actions to apprehend and remove non-criminal immigrants lacking authorization (Pew Research 2018).

Research in economics, demography, and sociology has widely documented the extensive costs of these deportations, including lost wages, food insecurity, mental distress, and home foreclosure (see Rugh and Hall 2016). In this paper we build on this work to evaluate how local efforts to deport immigrants altered residential racial segregation patterns across a broad range of US counties. To do so, we leverage variation in implementation of 287(g) agreements between local municipalities and the federal government during the 2000s, which gave police offers in affected areas substantial leeway to question and detain individuals suspected of being in violation of immigration law. We link these agreements to counties and assess change in patterns of residential segregation between Hispanics and other racial group members, by nesting census block data from the 2000 and 2010 decennial censuses within counties. We are able to identify the association between enforcement and segregation by comparing trends in segregation for counties that implemented 287(g) to a set of counties that applied for, but never implemented the program. To further isolate the effects, we are able to benchmark these estimates to corresponding changes in the segregation among non-Hispanic populations (e.g., between whites and blacks), which are unlikely to be affected by enforcement policy.

Results from difference-in-difference models indicate that 287(g) was associated with an increase in dissimilarity between whites and Hispanics of 2.2 points, stalling a secular decline in segregation between the groups by about 38%. (Alternative models comparing 287(g) counties to their contiguous neighbors produce similar, albeit weaker, estimates [1.3 point increase].) We find no measurable impact of 287(g) on Hispanic-black segregation, consistent with arguments that unauthorized Hispanics may find refuge in black neighborhoods (see Hall & Stringfield 2014). Importantly, we also find no impact of local enforcement on segregation between non-Hispanic groups.

In ongoing work, we use new estimates of local unauthorized populations to assess whether these associations are moderated by the size of populations likely to be targeted by enforcement policies. We are also exploring the possible mechanisms that may help explain how immigration policies bolster segregation, including the roles of deportation, selective-out migration, native prejudice, and ethnic attrition. Our analysis demonstrates how ostensibly neutral federal policies such as immigration enforcement interact with differential local engagement to produce a racially disparate impact that further segregates Latinos in the US, including citizens in mixed legal status households.