The launch of the Secure Communities program in 2008, and its subsequent gradual expansion to the entire country, constituted an important development in building up the crimmigration regime. The program, which requires that fingerprints gathered by any local law enforcement agency (LEA) at arrest be automatically checked against the Department of Homeland Security’s (DHS) IDENT database, has ignited considerable controversy. DHS officials and other proponents of the program have claimed that Secure Communities not only facilitates the detection of removable non-citizens, but also that it will improve public safety. Public safety benefits are expected, according to DHS, because automatic screening at an earlier point in the criminal justice process “reduce[s] the risk that an LEA will release a dangerous and removable criminal alien into the community.” Department of Homeland Security (DHS), Immigration and Customs Enforcement (ICE), Secure Communities: A Comprehensive Plan to Identify and Remove Criminal Aliens Strategic Plan 3 (2009). Opponents of the program cast doubt on any public safety dividends: on the contrary, they have predicted that marrying immigration enforcement to policing will alienate immigrant communities from law enforcement and reduce those communities’ willingness to report crimes and cooperate in investigations. Critics of the program also point to the risk of racial profiling and pretextual arrests. That is, agencies or officers with a preference for more aggressive immigration enforcement might be tempted to target Hispanic individuals for petty arrests, which otherwise would not be made, and even make arrests not supported by probable cause solely for the purposes of screening.
In a current working paper, Estimating the Effects of Immigration Enforcement on Local Policing and Crime: Evidence from the Secure Communities Program, Aaron Chalfin, Charles Loeffler, and I examine the empirical evidence for the effects of Secure Communities on public safety and policing in 335 large and mid-sized cities in 41 states. Both the proponents and critics of the program have suggested that there is some evidence for their views. With regard to public safety, DHS officials point to improved targeting of criminal aliens as evidence of the program’s effectiveness: they note, for example, that the share of criminal aliens among all removed non-citizens has been increasing under Secure Communities, and that the share of the pettiest offenders among the removed criminal aliens is shrinking. See Government Accountability Office (GAO), Report to the Ranking Member, Committee on Homeland Security, House of Representatives, “Secure Communities: Criminal Alien Removals Increased, but Technology Planning Improvements Needed” 20-21 (2012). However, these metrics have, at best, a tenuous relationship to public safety. As the program’s critics point out, the most common crimes of which non-citizens identified through Secure Communities were convicted remain minor.
To examine the effects of Secure Communities on criminal offending that actually threatens public safety, we examine the effects of activating Secure Communities on monthly rates of seven “index crimes” reported by the LEAs to the FBI between 2008 and 2011: murder, rape, robbery, burglary, assault, larceny, and motor vehicle theft. If Secure Communities successfully removed or deterred serious offenders, one might expect that in the months after its activation, the rates of more common crimes such as burglary, assault, and larceny would decrease. We control for national time trends and city-specific trends in these crime rates, and find no evidence that activating Secure Communities delivered a palpable boost to public safety.
We further examined subsets of our 335 cities, focusing on those cities with the most intense targeted enforcement since the activation of Secure Communities (i.e., highest removal rates of criminal aliens), those with the most intense overall enforcement (i.e., highest removal rates of all non-citizens), and those with the largest foreign-born populations. By and large, this analysis produces no robust evidence of public safety dividends. Places with the highest shares of foreign-born – that is, with the highest shares of populations whose behavior is most likely affected by the risk of deportation – did not experience statistically detectable reductions in crime after Secure Communities. The one exception to the general lack of significant effects is the finding that places with the most intense overall enforcement rates experienced a decline in assault rates after the program’s activation. Notably, no such consequences were discernible in the subset of cities with most targeted enforcement. In other words, places that have identified and removed the greatest numbers of criminal aliens (relative to the foreign-born population and adjusted for the time since the program’s activation) did not experience a decline in assault or other crime rates.
With regard to policing behavior, critics worry that easy access to immigration screening incentivizes the kind of behavior that has been overseen by Sheriffs Joe Arpaio of Maricopa County, Arizona and Terry Johnson of Alamance County, North Carolina. Acting unilaterally and under the auspices of 287g programs, these Sheriffs’ Offices were found to have engaged in unconstitutional targeting of Latino individuals. By attaching automatic immigration screening to mere arrests (rather than convictions), Secure Communities may lead officers to make arrests largely for the purposes of screening. We examine two possible manifestations of such effects. First, we examine trends in arrest rates for violent, property, and minor crimes before and after activation of the program. If police are influenced by the availability of automatic screening, they might be tempted to make more arrests for minor crimes, such as drug offenses and liquor law violations, with regard to which they exercise considerable discretion. However, we find no evidence that rates of arrests for any of the crime categories changed in the wake of Secure Communities, suggesting that police officers are not using Secure Communities to raise the stakes of minor crimes.
A more worrisome possibility is that police may become more likely to arrest Hispanic individuals relative to others, perhaps without altering the overall rate of arrests. Our ability to investigate this expectation is limited because most LEAs do not reliably report the ethnicity of those arrested. We investigated this possibility indirectly, by looking at the rates of arrests of White suspects relative to arrests of Black suspects, on the assumption that most Hispanic arrestees are typically recorded as White. Here too, we find no evidence that activating Secure Communities was followed by a statistically significant increase in arrests of White relative to Black suspects for any crime category, suggesting that critics’ fears of racial profiling are not being realized.
There is, of course, the possibility that not all officers are equally motivated to arrest potential immigration violators. In fact, many LEAs have stated their opposition to the program and concern that their mandatory involvement harms their relationship with their immigrant communities. Thus, it is probably not sensible to expect clear changes in policing practices across a broad set of LEAs. To explore the possibility that policing patterns changed in the subset of LEAs most likely to embrace involvement in immigration enforcement, we relied on county partisanship measures as a rough indicator of such attitudes. Several studies have found that jurisdictions that lean Republican are more likely to adopt aggressive measures aimed at unlawful immigrants, and those that lean Democratic are more likely to adopt “sanctuary” measures. We then look separately at LEAs serving majority Republican counties and those serving majority Democratic counties, and we find no evidence that arrest patterns were significantly affected by the activation of Secure Communities in either set of LEAs.
Overall, we find that the program has not augmented public safety during the first three or so years of its activity. We suspect that part of the explanation is that the most serious threats to public safety were adequately detected by prior programs, such as the Criminal Alien Program (CAP), and not many such individuals were slipping through the cracks. If correct, our results suggest that the wisdom of a policy involving local police in immigration enforcement in this way should be assessed on criteria other than public safety. Nor do we find evidence of marked changes in the behavior of local law enforcement.
A few thoughts as to why biased and pretextual policing may not be a very likely result of Secure Communities: as scholars and immigrants’ rights advocates have observed in the context of prior immigration enforcement programs involving local police, at least some LEAs altered their practices in response to the prospects of placing immigration violators into the deportation pipeline. It may well be that the LEAs most eager to catch immigration violators have already taken advantage of the previously available mechanisms for identifying and channeling non-citizens into removal proceedings. If so, then the marginal impact of an additional mechanism triggering immigration screening may simply be too faint to be detected, as there may be limited opportunities and high opportunity costs of further adjustments to police tactics. However, we do think that further research is needed – in particular, research focusing on a smaller set of LEAs and employing more finely-grained crime and policing data – before any firm conclusions about the program’s effects are drawn.
Elina Treyger is an Assistant Professor at the George Mason University School of Law; Aaron Chalfin is an Assistant Professor at the University of Cincinnati School of Criminal Justice; Charles Loeffler is an Assistant Professor in the Department of Criminology at the University of Pennsylvania.
Find this information useful? Then let others know about crImmigration.com, as well as my Twitter, Facebook, and LinkedIn pages. And to make sure you don’t miss an update, subscribe to the blog by entering your email address in the subscription box that appears on every page.