US Attorney General Jeff Sessions speaks at the federal courthouse in Central Islip, New York, on Friday. His speech was focused on gang violence and prosecution for illegal immigration by federal and local authorities. (EPA/Peter Foley)

In a speech Friday morning at the federal courthouse in Central Islip, N.Y., Attorney General Jeff Sessions made the following comment:

“The Bureau of Justice Statistics just released a report showing that 42 percent of defendants charged in U.S. district court were non-U.S. citizens.”

Sessions made the remark while discussing “so-called ‘sanctuary’ jurisdictions,” and district attorneys who give “special treatment to illegal aliens,” which he said “baffles me. “The  sanctuary jurisdictions “make our country less safe,” according to the attorney general, with the implication that disproportionate amounts of crime — 42 percent? — are committed by undocumented immigrants.

But the report Sessions referenced had another interesting statistic: Almost half of all arrests made by all federal agencies across the U.S. in 2014 were for immigration violations, largely along the border with Mexico. And of those who wound up charged in federal district court with criminal immigration violations, 92 percent were non-citizens, the Justice Department told us Monday. That will drive up the “non-citizen charges” number.

Further, of the 42 percent of non-citizens charged in district court, 72 percent were charged with an immigration offense, 17 percent with a drug crime and six percent with a property crime. One percent of the non-citizens prosecuted in district court were charged with a violent crime, according to the Bureau of Justice Statistics.

Now, most violent crime is prosecuted by the states, not the feds, so it’s not surprising that violent crime arrests are low at the federal level. For all defendants, violent crime only accounts for two percent of federal arrests. [Sex offenses, counted separately, also make up only two percent of federal arrests.] But Sessions’s comments raised again the question about a link between immigration and crime.

This is a subject criminologists have tackled with gusto recently. One new meta-study, by Charis E. Kubrin of the University of California at Irvine and Graham C. Ousey of the College of William & Mary, analyzed 51 studies done between 1994 and 2014 on the relationship between immigration and crime and found that, if anything, immigration reduces crime. “Where you have immigrants, you have less violent crime. Period,” Kubrin said Monday. The meta-study, to be published in the Annual Review of Criminology, even poses the question, “What role might immigration have played in America’s crime drop?” since crime levels have plummeted in the U.S. in the last two decades, and studies of that drop have focused mainly on policing, incarceration, guns and economics.

“You would be hard-pressed,” Kubrin said, “to find a criminologist to tell you that immigration and crime are related.” In looking at studies of the issue, “in this one area, it’s unbelievable the consistency with which this occurs.” The review of 51 scientific studies found that in reports with statistically significant size estimates, “the majority of the statistically significant results are negative, suggesting that greater immigration is associated with lower crime rates. In fact, our review indicates that significant negative effects [i.e. crime decreasing] are 2.5 times as common as significant positive effects. Taken alone, these descriptive results suggest a conclusion that rings familiar to many scholars — that immigration has a null or negative effect on crime rates.”

Kubrin was surprised to find that the Justice Department maintained statistics on race and country of national origin for federal defendants, which is typically not kept by local and state law enforcement agencies.

The report cited by Sessions, “Federal Justice Statistics 2013-2014,” by Justice statistician Mark Motivans, shows that 58.7 percent of people charged in U.S. district court in 2014 were from the U.S., 31.5 percent were from Mexico, five percent from Central America, and 0.2 percent from Canada. The racial breakdown of defendants in 2014 was 22.1 percent white, 19.4 percent black and 54.6 percent Hispanic. Motivans said the information came from interviews done by federal pretrial service officers to help determine whether a defendant should be released before trial.

Digging deeper, of the 165,295 arrests made by all federal agents in 2014, slightly more than 100,000 were made in the five federal districts along the U.S.-Mexico border, or about 61 percent. When the 81,881 immigration arrests in 2014 nationwide were analyzed, 93 percent were made in the same five border districts — southern California, Arizona, New Mexico, western Texas and southern Texas. The report notes that defendants with Mexican citizenship decreased from more than 25,000 in 2012 to about 20,000 in 2014, as did the number of defendants from Central America and South America.

But most immigration cases don’t go to the federal district court—which are the courts Sessions was discussing. And the immigration cases that do make it to federal court don’t account for most of the district judges’ caseload.

Only 30 percent of the more than 81,000 defendants arrested for immigration violations in 2014 were prosecuted in district courts.  The rest were handled as misdemeanors before federal magistrates, and mostly by magistrates concentrated in the five federal districts along the U.S-Mexico border. Magistrates at the southern border handled 65 percent of all immigration cases, with the other six percent fielded across the country.

Federal district judges saw more drug cases –which were 28.6 percent of all district court cases- than immigration cases (27 percent).  Property crimes cases were at 13.9 percent in that time frame and weapons crimes were at 7.9 percent.

Kubrin noted that discussing federal court numbers in trying to analyze violent crime, or associate it with a certain group, is fruitless. “It makes up such a small percentage of offending in the U.S.,” she said. “It’s studying where the action isn’t.”

Here is the report Sessions cited: