“When cities like Philadelphia, Boston, or San Francisco advertise that they have these policies, the criminals take notice. According to a recent study from the University of California Riverside, cities with these policies have more violent crime on average than those that don’t.”
— Attorney General Jeff Sessions, speech on sanctuary cities in Las Vegas, July 12, 2017
In a speech about “sanctuary cities,” Sessions cited research from University of California at Riverside that does not actually support his point. There is little research looking at the impact of sanctuary policies on crime. It’s a difficult correlation to study; many factors affect crime, and state and local law enforcement do not always track inmate citizenship status.
We previously dug into the existing research on sanctuary cities and crime, including the one he mentioned, and did not find that criminals “take notice” when cities have sanctuary policies. Sanctuary jurisdictions release inmates after their criminal case is complete, and extensive research shows noncitizens are not more prone to criminality than U.S.-born citizens. Moreover, some sanctuary jurisdictions do cooperate with the federal government if they believe the inmate is a public safety threat.
What is Sessions talking about?
There’s no official definition of “sanctuary,” but it generally refers to rules restricting state and local governments from alerting federal authorities about people who may be in the country illegally. Immigration enforcement is a federal responsibility, and state and local law enforcement can decide how much they want to cooperate with the federal government for immigration enforcement.
Immigration and Customs Enforcement (ICE) can issue an “immigration detainer,” or a request to be notified when a noncitizen convicted of a crime is being released at state or local levels. ICE can then take custody of the offenders and figure out if they should be deported.
But some agencies decide not to alert ICE. Some fear that victims and potential witnesses may not come forward to report crimes if they are afraid of being reported to federal authorities for their immigration status. Others cite dwindling local and state law enforcement resources.
There are somewhere between 165 and 608 local and state government with “sanctuary” policies. But the jurisdictions differ in their approach. Some don’t cooperate at all, while others only do so in civil investigations, for felony convictions or for offenders otherwise deemed to be a public safety threat.
The August 2016 study by researchers at UC Riverside and Highlight College looked at about 80 jurisdictions and used FBI city-level crime data to see how violent and property crime rates changed after sanctuary policies were adopted. Then they compared each sanctuary city to a similarly situated, non-sanctuary city, based on census data and other variables.
According to the Justice Department, Sessions is citing this graphic in the study that shows that violent crime rates in cities with sanctuary policies are higher than violent crime rates in cities without sanctuary policies:
Sessions is presenting the data accurately, and making a “different conclusion based on the data as well as his experience and discussions with law enforcement officials throughout the country,” according to Justice Department.
But the researchers used this graphic to show that there is no correlation between violent crime and sanctuary policies. Here’s what they said in their research about this graphic:
“If the two lines for each year cross each other at any point, then the relationship between violent crime and city-type is not statistically significant. While there is a mild tendency to have slightly more crime in sanctuary cities, these effects are very small, and are not statistically significant. In general, the crime rate per 100,000 people differs between 100-200 incidents a year; however, again, these effects remain statistically insignificant.”
In October 2016, the authors summarized their research in Monkey Cage, a blog hosted by Washingtonpost.com. The authors acknowledged in that column that violent crime is slightly higher in sanctuary cities than non-sanctuary cities, but again, they warned about the statistic significance: “However, judging by the error bands — which capture the uncertainty underlying these estimates — the relationship is not statistically significant before or after a sanctuary policy is passed. We find similar results for property crime and rape.”
They added: “That is, a sanctuary policy itself has no statistically meaningful effect on crime.”
Researchers ran several other tests to find whether there was connection. One graph shows that there is no clear generalizable pattern about violent crime and whether the city has sanctuary policies or not.
After Sessions’s speech, the researchers wrote another column for Monkey Cage disputing the attorney general’s use of their data. One of the researchers, Loren Collingwood, assistant professor at UC Riverside, told The Fact Checker: “Our analysis, findings, and article are based on rigorous statistical methods, which are part and parcel of the interpretation. With academic research you must consider the point estimates and level of uncertainty (confidence bands) when making interpretations. Same thing should apply when formulating national policy.”
But the Justice Department said the authors are arguing about the statistical significance of the data, but that Sessions is citing what the data show, not the statistical significance.
“The Attorney General accurately cited data from a study that clearly showed that the violent crime rate was higher in sanctuary cities versus non-sanctuary cities. That is a fact and assertions to the contrary ignore that data,” said Ian Prior, Justice Department spokesman.
The Pinocchio Test
Sessions is being rather misleading by citing data from a study that doesn’t support his point.
Sessions says that “criminals take notice” when cities make it known that they have “sanctuary” policies that restrict local cooperation with federal authorities about people who may be in the country illegally. He cites data from a study by the University of California Riverside, to say that “cities with these policies have more violent crime on average than those that don’t.”
But he omits the other side of the research, which is that this data point is not statistically significant. The study by the researchers did not find that sanctuary policies had any effect on crime — a point that they have emphasized in the past and which Sessions appears to have willfully ignored for political purposes. We award Sessions Three Pinocchios for twisting the data out of context to make the opposite point that the researchers made.
Send us facts to check by filling out this form
Keep tabs on Trump’s promises with our Trump Promise Tracker
Sign up for The Fact Checker weekly newsletter