From Pre-Text Stops to Immigration Detention: State Immigration Policies and County Jail Collaboration with ICE from 2014-2018

Background/Purpose: Punitive anti-immigrant policies increasing local law enforcement’s (LEs) enforcement and detention capabilities and communication with Immigration and Customs Enforcement (ICE) have proliferated among state and local governments since the mid-2000s. Research has suggested that such policies amplify LE’s capacity to collaborate with ICE post-arrest, thereby exacerbating the possibility of deportation after arrest. Nevertheless, literature has yet to substantively examine a critical element of LEs’ utility for ICE: their widespread discretion to stop and arrest. Policies reducing LEs’ discretion to engage in pre-text stops and demand identification during routine encounters may indirectly protect migrants by limiting the passive information sharing between LEs and ICE. Laws that criminalize actions or behaviors based on individuals’ legal status may enhance migrants’ vulnerability to arrest. Thus, I use mixed-effects zero-inflated negative binomial regression models (ZINB) to test the relationship between state-level policies heightening LE’s power pre-arrest and post-arrest collaboration with ICE on the number of migrants detained in county jails from 2014 to 2018.

Methods: I use county-level panel data from numerous secondary data sources including Vera Institute of Justice, ICE, and National Conference of State Legislatures (N=11,298). The dependent variable is the count of migrants held in county jails for ICE. I operationalize pre-arrest policies with 2 variables: the presence of a stop and identify law and the number policies increasing migrants’ vulnerability to arrest. The final independent variables measured the number of state-level post-arrest communication and detainer collaboration policies. Covariates include urbanicity, percent-Latine residents, excess jail bedspace, jail capacity, unemployment, political majority (presidential), time, and state or local level sanctuary policies. Overdispersion and intra-class correlation tests indicated that mixed-effects ZINB (with state-level random effects) was appropriate.

Results: Each increase in pre-arrest policies increased the expected count of immigrants held for ICE by 11.97% (B=.11, SE=.06, p=.04). The number of migrants held for ICE increased from 45 to 79 as states moved from 0 to 5 pre-arrest policies. Stop and identify laws increased the expected count by 104.33% (B=.71, SE=.33, p=.03), from 45 for counties in states without such policies to 96 for counties with them. However, policies focused on LEs’ communication and detainer collaboration with ICE post-arrest did not significantly affect the number of migrants county jails held for ICE.

Conclusions/Implications: My findings demonstrate the need for research that critically evaluates the nuances state-level policies affecting LEs’ utility to ICE, both before and after arrest. While research has frequently collapsed these policy domains, my findings suggest that pre-arrest policies may affect LEs’ holds for ICE more than policies shaping LEs’ communication and collaboration with ICE post-arrest. Insofar as research and advocacy ignores the consequences of policies heightening LEs’ arrest power and surveillance capacity, it leaves the human cost of these structures unquestioned. Social work policy advocacy must advance local, state, and federal reforms that support migrants’ constitutional and human rights and recognizing LEs’ power to stop and arrest as foundational to the violence induced through the immigration enforcement and criminal legal systems.