Sanctuary Policies and City-Level Incidents of Violence, 1990 to 2010

Abstract

Despite media coverage of isolated incidents of violent crime perpetuated by undocumented immigrants in cities with sanctuary policies, there is scant systematic research on the relationships between the adoption of sanctuary policies, unauthorized immigration, and crime. We compile city-level data from official sources and use fixed-effects negative binomial regression to examine whether the adoption of city-level sanctuary policies and the concentration of unauthorized Mexican immigrants are associated with homicide and robbery incidents in 107 U.S. cities, across three decades. We find evidence that the adoption of sanctuary policies is associated with a reduction in robberies but not homicide. In contrast, an increase in the relative size of a city’s unauthorized Mexican immigrant population corresponds with a reduction in homicide; however, only in sanctuary cities. Lastly, shifts in violence during our study period are consistently related to social structural characteristics of cities, which are findings consistent with social disorganization theory.

Keywords:

• sanctuary cities
• homicide
• robbery
• social disorganization
• undocumented immigrants
• Latino immigration

Acknowledgments

We would like to thank Christina Diaz, Jeremy Fiel, Karam Hwang, David Rigby, Charles Seguin, and two anonymous reviewers for their suggestions and comments on earlier versions of this research.

Notes

¹ We acknowledge that the term “sanctuary”, which is often used to describe policies or jurisdictions that limit the involvement of local law enforcement in enforcing federal immigration law, is contested and characterized by some as a misnomer since these policies or jurisdictions do not protect migrants from federal enforcement efforts (Tramonte, 2011). However, given its colloquial use in political and popular discourse, we use the term “sanctuary” throughout this paper for the sake of clarity and parsimony.

² Contrary to these arguments, data from the Criminal Alien Program (2010–2013) suggest that the majority of “criminal aliens” were removed for non-serious offenses related to violations of traffic, drug, or immigration laws, not serious violent offenses (Cantor, Noferi, & Martinez, 2015). Regardless, many “sanctuary” policies, such as San Francisco’s, allow for local government assistance and cooperation in the enforcement of federal immigration law when individuals have been convicted of violent or serious felony offenses (see San Francisco’s Due Process for All and Sanctuary 96-16 ordinance). Thus, sanctuary cities do not completely protect non-citizens from deportation (Tramonte, 2011).

³ See Bau (1994) for a discussion of the 1980s Sanctuary Movement.

⁴ See San Francisco’s Due Process for All and Sanctuary 96-16 ordinance for common exceptions.

⁵ Kittrie (2006) notes that other, though less publicized, justifications include the belief that local law enforcement officials do not have the resources, authority, or proper training necessary for the enforcement of immigration laws in cooperation with the federal government. Proper training was noted to be especially important to ensure that police officers would avoid racial profiling.

⁶ See http://www.ailadownloads.org/advo/NILC-LocalLawsResolutionsAndPoliciesLimitingImmEnforcement.pdf.

⁷ See for example https://cis.org/Map-Sanctuary-Cities-Counties-and-States.

⁸ See http://ojjpac.org/sanctuary.asp.

⁹ Like Ousey and Kubrin (2014) we constructed an additional disadvantage index that excluded percent black. The alternative construction did not significantly alter the results of our analyses.

¹⁰ These results may be expected by routine activities theory (Cohen & Felson, 1979). When a larger proportion of the population is “high income,” there is a greater (relative) size of suitable robbery targets at the city-level. Furthermore, robberies (or other crimes like burglaries) can “go wrong” leading to robbery-turned-homicide. This could explain a relationship with both measures, but a systematic investigation of this is beyond the scope of the current paper.

¹¹ Possibly because of language barriers or homophilic preferences.

Additional information

Notes on contributors

Ricardo D. Martínez-Schuldt

Ricardo Martínez-Schuldt is a doctoral candidate in the Department of Sociology at UNC-Chapel Hill. His research focuses how local contexts impact the needs of immigrant populations and how sending-states facilitate the integration of migrants across borders. Additionally, he conducts research on neighborhood and city-level correlates of crime, victimization, and reporting.

Daniel E. Martínez

Daniel E. Martínez is an assistant professor in the School of Sociology at the University of Arizona. He is a co-principal investigator of the Migrant Border Crossing Study, which is a binational research project focusing on unauthorized migrants’ border crossing, apprehension, and repatriation experiences.