On Feb. 10 of last year, a Justice Department lawyer in the department’s National Security Division (NSD) assembled some data on international terrorism convictions for transmission to the White House. The lawyer, a man named George Toscas, included in his email to his superiors what he described as “some general statements that are supported by [the data] and can be used publicly.”
They included such anodyne claims as these:
- “Since 9/11, convictions have been obtained against hundreds of defendants for terrorism or terrorism-related charges in Article III courts”; and
- “We have a long history of using the criminal justice system to incapacitate individuals who pose a threat to the U.S. and its interests here and abroad. Since 9/11, hundreds of convictions have been obtained in our federal courts.”
A few days later, Justice Department officials had a back-and-forth with the FBI in which they tried to answer a strange question on which they didn’t keep data, apparently in response to a query from departmental higher-ups or from the White House itself: Where were the people described in the earlier data set born?
On Feb. 28, President Donald Trump stood before a joint sessions of Congress and gave a speech in which he characterized the data Toscas and the Justice Department had assembled. He did not use any of Toscas’s careful statements—or anything remotely like them.
Here’s what he said: “According to data provided by the Department of Justice, the vast majority of individuals convicted of terrorism and terrorism-related offense since 9/11 came here from outside of our country.”
When Trump made his remarkable statement, we did not believe it, and we set out to evaluate its accuracy—both as to the role of people born abroad in terrorism incidents and as to the Justice Department data on the subject. Now we have documents—some of the Justice Department correspondence that preceded Trump’s speech, released to us last week under the Freedom of Information Act.
In April 2017, Lisa Daniels and Nora Ellingsen analyzed the prior version of the list which Toscas sent his superiors, which included international terrorism or terrorism related convictions through the end of 2015. Part I, Part II, and Part III of that analysis were published on Lawfare. That evaluation indicated that the administration’s statements regarding the role of foreign-born individuals in the United States’ terrorism problem were inaccurate. Ellingsen and Daniels determined: “The data Trump cited in his speech to the joint session of Congress simply don’t support his claims that a ‘vast majority’ of individuals on the list came from outside the United States—unless, that is, you include individuals who were forcibly brought to the United States in order to be prosecuted and exclude all domestic terrorism cases.” If you count reasonably, Ellingsen and Daniels concluded, instead of accounting for between 63 and 71 percent of terrorism convictions, foreign-born persons would likely account for only 18 to 21 percent of terrorism convictions.”
In the wake of that analysis, Benjamin Wittes filed a Freedom of Information Act suit to find out how precisely the president came to so dramatically mischaracterize Justice Department data by way of imputing a disproportionate share of terrorism to immigrants. As he described the underlying FOIA request:
I'm going to be very blunt here: I not only believe that the White House made up "alternative facts" about the substance of this matter in a presidential address to a Joint Session of Congress, I don't believe that the National Security Division of the Justice Department provided any data or analysis to the White House that could reasonably be read to support the president's claim. In other words, I believe the president was lying not merely about the underlying facts but also about his own Justice Department. Or, in the alternative, I believe it's possible that the Office of the Attorney General may have supported the White House's claim. But I think it extraordinarily unlikely that the folks at NSD actually provided data in support of this presidential statement.
Here's why I believe this: I know a lot of people at NSD, and they are not the sort of people who grossly mischaracterize facts in order to make political points. Indeed, I believe that the folks there have the integrity to raise internally the very issues that Ellingsen and Daniels raised in these pieces.
This week, we received 57 pages of material from the National Security Division in response to this lawsuit. All of it is available here:
The material is sketchy and does not answer all of our questions; the litigation is thus continuing. Moreover, perhaps the key document, as we shall describe, has been withheld. But it is safe to conclude even from the limited material we were given that career officials at the Justice Department did not, in fact, support the irresponsible and false statement the president made. Indeed, while they provided the updated list of cases, they specifically warned that the material excluded domestic terrorism matters and could only be used to support generic statements that the Justice Department had prosecuted a lot of cases. And while they worked with FBI to scrape together some national origin information, they did so only with respect to that dataset, and the FBI specifically warned that it might not be accurate.
Here’s what we can reconstruct. At some point early in its tenure, the Trump administration asked NSD to provide data on international terrorism prosecutions and where the perpetrators came from. It’s unclear precisely where this request came from, but it appears to have been transmitted through the Office of the Attorney General. The request may have been in support of the president’s upcoming address to the joint session of Congress or it may have been in connection with the ongoing processes associated with the travel ban executive orders.
In response, the Justice Department updated NSD’s chart of public and unsealed international terrorism and terrorism-related convictions—the same chart that Ellingsen and Daniels analyzed back in April—to include convictions through Dec. 31, 2016. Emails note that the Justice Department completed that update on Feb. 10, 2017. The new updated chart is available as part of the document release.
The updated chart, which includes 40 new international terrorism convictions from 2016, is not only consistent with the Ellingsen-Daniels analysis, but further undercuts President Trump’s statements. Of the defendants convicted in 2016, a minority—less than 43 percent—were foreign born. Furthermore, only six defendants were born in countries included in the latest iteration of the administration’s travel ban—a scant 15 percent of the 40 new defendants. Two defendants immigrated to the United States from Somalia, two from Sudan, one from Iraq, and one from Yemen. Finally, extraditions again account for a significant proportion of the foreign born defendants. In 2016, almost 18 percent of foreign-born convictees were extradited to the United States from overseas.
But the Justice Department’s data didn’t quite answer the White House’s question. The Justice Department actually did not keep data on where people convicted of international terrorism or terrorism-related crimes are born. Naturalized citizens, after all, are citizens, and while immigration status may matter, place of birth does not. The result was that the data the White House was seeking didn’t exist in one place.
So on Feb. 10, Tashina Gauhar, an official in the deputy attorney general’s office, emailed the updated list to a woman in the attorney general’s office named Rachel Tucker. She forwarded the data prepared by Toscas, saying that NSD was comfortable “with your sending this chart to the WH and/or DHS as needed to get the additional information [on] the stats we previously discussed”—presumably a reference to the place of birth matter on which NSD didn’t have data. She also forwarded a previous message from Toscas in which he described what work he thought the data could reasonably do. In was in that email that Toscas writes the responsible model sentences, quoted above.
In addition, Toscas wrote that the “scope and limitations of the information contained in the chart are fully described in the preamble, which should always accompany the chart.” In the same Feb. 10 email, he warned: “IMPORTANT FYI: We do not use the total (600+) number of convictions listed on the chart in any statements because the explanations/caveats in the preamble always need to accompany that total number--which is why the chart and the preamble should always be distributed and read together.”
What are those explanations and caveats? The preamble itself is close to two pages long, and contains several additional warnings—most notably that the chart “does not include convictions related solely to domestic terrorism.” As Daniels and Ellingsen noted in April 2017, “domestic attacks—cases in which there appears to be no foreign nexus—have accounted for the majority of terrorism offenses in the U.S.”
In order to track down the national-origin information, the Justice Department tasked the FBI to do a national origin analysis on 553 convictions (excluding the 115 convictions arising from the nationwide investigation conducted after September 11th). On Feb. 22, 2017, James Rybicki, then-Director James Comey’s chief of staff, sent an email to James Crowell and Tashina Gauhar in the deputy attorney general’s office with a rough cut of the data. Rybicki explained that the data had been a manual pull by analysts in the FBI’s Counterterrorism Division, and he too provided a cautionary note from the division:
Given the use of aliases in IT matters, factual errors in the initial data, and conflicting DOBs, database checks are limited in their ability to accurately identify a date/place of birth. We highlight these concerns to note the attached list likely contains gaps or error, although we endeavored to provide the most accurate data as rapidly as possible. This spreadsheet contains the unclassified version of the data that mevoces the FBI case file.
The FBI chart was forward by George Toscas in the National Security Division to Rachael Tucker and Jody Hunt in attorney general’s office on Feb. 27, 2017 at 10:30 p.m.—the night before President Trump’s address.
What’s left out of the Justice Department’s production is a classified email sent on Feb. 14, 2017 by Mary McCord to the Office of Legal Counsel. The document is two pages long: “I will also send you on TS the email I sent to Rosemary after consulting with the Bureau,” McCord wrote in a separate email, with the subject line “FW: Draft Findings.” McCord was then acting head of NSD, and we suspect—although we don’t know for sure—that the email withheld from us reflects concerns on her part along the lines of the Ellingsen-Daniels analysis.
In short, the NSD documents reveal, sketchy though they are, that in response to the White House’s requests, NSD produced a list of international terrorism and terrorism-related convictions, which the career prosecutors specifically advised could be used only to make general statements about the robustness of the criminal justice system as a counterterrorism tool and which they specifically noted did not include domestic terrorism cases.
When asked to produce national-origin data about those on the list, the FBI generated the data manually but specifically warned that their quick and dirty effort “likely contains gaps or error.” And the head of NSD sent a two-page email that may well contain more specific warnings.
The White House took all this material and crafted out of it a statement, given by the President personally in a joint session, in which the word “international” was removed, thus rendering the entire statement false by wrongly including within the statement all the domestic terrorism cases NSD had specifically noted were not included, and in which none of NSD’s caveats appear. As recently as last month, President Trump repeated this untruth in response to a new DOJ-DHS report:
....we need to keep America safe, including moving away from a random chain migration and lottery system, to one that is merit-based. https://t.co/7PtoSFK1n2
— Donald J. Trump (@realDonaldTrump) January 16, 2018
We have not yet received any production reflecting the conduct in this matter by either the White House or the attorney general’s office—so it’s not clear at this stage whether the White House actively misrepresented NSD data or whether Attorney General Sessions’s staff did the misrepresentation and the president merely parroted the attorney general. The litigation will continue until we find that out.