This term the U.S Supreme Court will hear argument in Jennings v. Rodriguez revisiting, for the first time in over a decade, constitutional concerns with mandatory detention of noncitizens facing deportation. The last time the Court considered this issue, in Demore v. Kim, it relied heavily on government-proffered data to conclude that immigration detention was sufficiently brief to avoid constitutional concerns regarding the associated deprivation of liberty.

The problem with the government’s data in Demore is that it was wrong, and immigration detention is far from brief. InDemore, the government claimed that the average stay in immigration detention for affected individuals that appeal their case was about four months. Yet, this summer the Solicitor General’s Office conceded that the actual average was 382 days (more than a year).

Given the central role that this data point played in the Court’s analysis in Demore, one would hope the government would be particularly careful in its representations in Jennings. Yet, in its brief, the government mischaracterizes and relies on faulty data to marshal its arguments for why prolonged mandatory detention—without any individualized determination of danger to the community or flight risk — is defensible.

The Solicitor General cites to a Boston Globe report that allegedly found a high criminal recidivism rate among individuals with criminal records who are released from ICE custody. The Solicitor General states:

Similarly, a recent study calculated a recidivism rate of approximately 30% among 323 criminal aliens released from immigration custody in New England, for any reason, from 2008 to 2012. Maria Sacchetti, Criminal Aliens Reoffend at Higher Rates than ICE Has Suggested, Boston Globe, June 4, 2016.

Gov’t Br., at 33. (emphasis added).

The Solicitor General’s reliance on the Globe “study” is problematic for several reasons.

First, as the Globe conceded—and the Solicitor General surely should have known—the data pertained only to detainees with prior criminal convictions who were released based on the Supreme Court’s decision in Zadvydas. (Boston Globe case timeline, entry for Dec. 22, 2011). Under Zadvydas, the government is not permitted to continue detaining noncitizens who have been ordered removed but whose country won’t take them back. Instead, with minor exceptions, the government must release such individuals under conditions of supervision—without a bond hearing or any individualized determination of danger or flight risk. In other words, by framing this data point in Jennings as pertaining to detainees who are released “for any reason,” the government ignores a fundamental difference between those who would be released under Zadvydas and those who should be protected by the rule at issue in Jennings. The rule in Jennings allows for release of individuals only if the government cannot prove to an immigration judge that they are a flight risk or a danger to the community, whereas the rule in Zadvydas requires release, without any determination of danger or flight risk, if removal is not reasonably foreseeable. Thus, the Solicitor General’s suggestion that the Globe’s recidivism data has any relevance for recidivism rates of individuals released underJennings is dubious at best.

Second, the Solicitor General fails to cite to Immigration and Customs Enforcement’s (ICE) own criminal recidivism data that undermines the government’s position. NIJC recently obtained through the Freedom of Information Act an ICE document for FY2013, which calculated a recidivism rate of less than 3% for all individuals released from ICE custody who had any previous criminal conviction. Why is the Solicitor General citing the Boston Globe’s 30% recidivism figure, yet ignoring its own more relevant data which contradicts that result?

Finally, the Globe “study” suffers from significant methodological flaws. In an email exchange between NIJC and the author, the journalist explained to NIJC some of the limitations of her data. For example, while her article chronicles a number of particularly heinous crimes committed by individuals released under Zadvydas, the author could not confirm whether these crimes were representative of the crimes she reported as being committed by 30% of the released detainees. Indeed, she never tabulated or published the range of crimes that the “recidivists” had committed.

These data limitations are critically important to whether the Supreme Court should lend any probative value to the Boston Globe’s recidivism figures. For example, routine traffic offenses and minor probation violations are technically recidivist offenses, yet hardly seem like the type of risks to public safety that would weigh in favor of mandatory detention without six month review for an entire category of individuals—a class of individuals who have already served their time for any previous convictions.

The U.S. Government should be in the best position to accurately describe, through data, the inner workings of the U.S. immigration system, particularly as it relates to detention. Yet, time and again, it has misrepresented reality in such a way that should cause the Court to seriously question the government’s representations about how things “work” within its own system. And in light of these problems, the Court should view the government’s data with appropriate skepticism, lest we have another decade of punitive over-detention based on faulty data and misrepresentations.