The Transactional Records Access Clearinghouse at Syracuse University released a new report entitled Federal Agencies Can’t Agree on Who is a Terrorist. TRAC always provides great information, and is an invaluable resource for counterterrorism scholars looking for data to support their arguments.
With that said, I don’t agree with the conclusions in this report and the heavy focus placed on the declination rate in terrorism cases. The declination rate is the percentage of cases referred to prosecutors by investigators (usually the FBI) in which the U.S. attorney declines prosecution. The report specifically highlights the fact that in two out of every three cases brought to prosecutors by investigative agencies, the prosecutors declined to bring charges. The report further notes that “[e]qually disturbing, rather than seeing this very high turndown rate diminish with time, the “declination” of terrorism referrals has been moving in the opposite direction — from 31% in FY 2002 to 61% in FY 2005 and to 73% in FY 2008.”
I’m less concerned by this declination rate than the authors of this report. In fact as the FBI increasingly orients its activites toward collection of intelligence and recruitment of informants, I’d expect the declination rate to rise. Moreover, it’s not clear to me that the declination rate supports the conclusion that there are fundamental definitional problems in CT cases. There are very good reasons for declinations, and the TRAC data states one of them: in 22.2% of cases prosecutors declined to prosecute due to “Weak or insufficient admissible evidence.” In most CT cases, the difficult issue faced by a prosecutor is determining whether information the government knows (and may very reliably know) can be used in court; if a prosecutor sees that an agent is relying in part on HUMINT , or a specific collection method, or ORCON non-dissemination information and no alternative admissible evidence is possible they’ll likely decline to prosecute. This is the type of pre-charge decision that CT prosecutors face everyday and the data TRAC relies on will show this as a declination, but won’t show the dynamics at work. The 22.2% number isn’t surprising in light of the realities of a national security investigation premised on intelligence information, combine that with the fact that many suspects referred to a prosecutor will subsequently become cooperating witnesses and the number can drop further. Definitional ambiguity may be part of this problem, but the realities of a national security investigation explain more to me than disputes over definitions do.
The other top reasons for declination don’t trouble me much either. If a prosecutor decides to push a case despite seeing potential difficulties with the admissibility of intelligence information, the data shows that in 14.1% of the cases there is an agency request for declination. That could be a case where the FBI and the prosecutor both think the case should be prosecuted but someone in the intelligence community requests a declination due to intelligence equities. Those big percentage declinations (22.2% and 14.1% respectively) when combined with the 5.2% office policy declination rate, and 4.9% prosecution by other authorities declinations tell me that quite a few declinations (nearly half) can be attributed to the complex interagency realities of CT prosecutions, not fundamental definitional problems.