Cross-posted on Lawfare
Our friend Carrie Cordero has levied criticisms against three of the recommendations presented in our report, What Went Wrong With the FISA Court. We appreciate, as always, her constructive engagement with us on these issues. In the same spirit, we offer these points in response.
Our report notes that Section 702 had a limited effect on the government’s ability to collect communications between foreigners without obtaining a FISA order; instead, its primary purpose and effect was to remove the requirement of an individualized order for the acquisition of communications between foreign targets and U.S. persons. Carrie responds that Section 702 was necessary because, under FISA, “the pre-2008 definitions in FISA technically required that the government obtain a probable-cause order from the Court in order to collect the communications of Terrorist A in Afghanistan with Terrorist B in Iraq.” But, as we explain in our report, this was true only for one category of foreign-to-foreign communications: e-mails stored on U.S. servers.
That’s because of the statute’s complicated definition of “electronic surveillance,” which is the activity that FISA regulates. The definition is broken down into three types of surveillance: acquisition of wire communications (which includes phone calls or Internet communications in transit over cables), acquisition of radio communications (which includes calls or Internet communications in transit through wireless means), and “monitoring” (which previously meant planting a bug, but today includes acquiring stored e-mails). For the first two categories, acquisition is defined as “electronic surveillance”only if one or more of the communicants is a U.S. person. In other words, for wire or radio communications between foreigners, 1978 FISA simply had nothing to say; “monitoring” is the only category of foreign-to-foreign communication that 1978 FISA regulated.
That’s why our report states that, “[w]ith the exception of e-mails stored in the United States, the new law had no impact on the government’s ability to collect the communications of foreigners with other foreigners.” Carrie is, of course, correct that Section 702 allows the government to obtain other types of foreign-to-foreign communications without a court order… but so did 1978 FISA, so Section 702 made no change there. The most significant change Section 702 made was to permit the acquisition of communications between foreign targets and U.S. persons without a court order. Carrie does not explain why this was necessary to allow the government to collect the communications of Terrorist A in Afghanistan with Terrorist B in Iraq.
Carrie also appears to interpret our recommendations as requiring a probable cause order whenever the government seeks to collect on a foreigner overseas. This isn’t what we’re proposing. Our recommendation is to repeal Section 702 “and replace it with a regime requiring an individualized court order for the interception of communications involving U.S. persons.”
We’re aware that differentiating between U.S. persons and foreigners is currently more difficult for stored e-mails than for communications obtained in transit, where relevant information may be gleaned from packet headers. That may support a more nuanced and creative solution for that category of surveillance. But there’s no conceivable reason why the solution should be programmatic acquisition of any type of communication sent by a foreign target.
In our report, we recommended establishing a Special Advocate to help address the FISA Court’s Article III shortcomings. We argued that it was insufficient to give FISA Court judges the option of soliciting amicus briefs, since they’ve always had this option and have chosen not to exercise it. Carrie suggests that perhaps the judges didn’t feel that more adversarial proceedings were necessary, and that may well be true (although some FISA Court judges have weighed in with their assessment that more adversariality would be quite helpful). But judges’ personal feelings about the value added by the adversarial process are not dispositive of what Article III requires.
Leaving aside the constitutional issues, Carrie argues that adding adversariality to FISA proceedings would be incongruous for three reasons.
- We would be providing greater ability to challenge surveillance to foreign intelligence targets, than to subjects of criminal investigation domestically, for whom wiretap applications are approved in camera ex parteby district court judges and magistrates.
In fact, foreign intelligence targets would have no ability to challenge surveillance under our Special Advocate proposal. They would not be able to choose and hire an attorney to present their case, and to communicate to this attorney the facts only they could know. Instead, they would remain ignorant of the surveillance, and a Special Advocate would do his/her best to present a case without their help or involvement. By contrast, subjects of searches in criminal cases are informed of the search. At a minimum, then, they can bring a civil suit, and they often have the opportunity to challenge the search as part of a criminal trial (an opportunity that’s rare in foreign intelligence cases). They can do so with full knowledge of their own actions and with an attorney of their choosing who works only for them. When it comes to the ability to challenge surveillance, the advantage clearly belongs to the subject of the criminal search, even under our proposal.
- We would be endorsing the concept of litigatingsurveillance conducted for national security purposesbefore it takes place, placing intelligence operations and national security at risk.
We assume Carrie is referring to the delay that “litigation” might involve. Certainly, if programmatic surveillance continues, there is nothing incongruous about conducting an adversarial proceeding on a proposal to collect millions of Americans’ information over a period of years. Whatever delay may occur is plainly justified under those circumstances, and it is far more incongruous to determine the legality of such a sweeping program without an adversary present.
If programmatic surveillance ends (as we recommend), there is still no reason to fear that time-sensitive investigations will be jeopardized by the participation of a Special Advocate. There is already a generous provision for emergencies in FISA, which allows the government to begin surveillance before even submitting an application to the FISA Court and to continue that surveillance pending the court’s ruling. The Justice Department need only certify that there’s an emergency situation and that the surveillance meets the necessary legal criteria. This provision would not be disturbed under our proposal.
- We would be creating an entire new bureaucracy duplicating the legal and oversight functions that the Justice Department, under the leadership of the Attorney General, and the Court, comprised of independent federal district court judges, perform.
To be clear, we’re not proposing the establishment of an office to conduct “legal oversight.” We’re proposing the establishment of a Special Advocate to advocate against the government’s position in court. We’re fairly certain that no one in the Justice Department or the FISA Court is charged with vigorously opposing the arguments presented by Justice Department attorneys in FISA Court proceedings, so “duplication” is not an issue here.
Carrie particularly disagrees with our recommendation that FISA surveillance should be available only if obtaining foreign intelligence is the primary purpose of collection. She equates this recommendation with a proposal to “rebuild ‘the wall.’” She suggests that “the wall” had catastrophic effects, citing a statement by prosecutor Patrick Fitzgerald to the effect that he was not allowed even to speak with intelligence investigators at the FBI while he was building a criminal case against Osama bin Laden.
Carrie ignores the legal issue at the center of this recommendation. Although the FISCR takes a different view, the circuit courts that have addressed this issue have held that the “primary purpose” test is necessary to trigger the foreign intelligence exception to the warrant requirement. Without such a requirement, these courts observed, it would be far too easy for the government to avoid the warrant requirement in ordinary criminal cases. Their concerns have come to pass, as the PCLOB reports that the FBI “with some frequency” searches databases containing Section 702 data when opening criminal investigations or assessments “unrelated to national security efforts.”
Critics of reform posit dire consequences to reinstating this vital Fourth Amendment protection. Our report identified three flaws in the logic chain that purports to lead from the “primary purpose” test to disaster, and Carrie’s argument tracks two of them precisely. First, our report notes that nothing in the “primary purpose” test necessitated the particular limits that the Justice Department imposed on itself. Carrie doesn’t dispute this point; instead, she states, “the reality is that the rules and restrictions wereput in place.” That’s a truism; it doesn’t suggest that the Justice Department would follow exactly the same course today if a “primary purpose” test were reinstated. We strongly suspect the Department would take a different approach in operationalizing this requirement.
Second, the problems that followed from the limits the Justice Department imposed on itself do not appear to have resulted from the rules themselves – the FISA Court described these rules as permitting “broad information sharing” and “substantial consultation and coordination” – but from a widespread misunderstanding of what they required. Indeed, the 9/11 Commission described the term “the wall” as “misleading,” and notes that the rules in question “were almost immediately misunderstood and misapplied.”
Fitzgerald’s quote illustrates this point nicely. He says,
But there was one group of people we were not permitted to talk to. Who? The FBI agents across the street from us in lower Manhattan assigned to a parallel intelligence investigation of Usama Bin Laden and al Qaeda. We could not learn what information they had gathered.
That isn’t a remotely accurate description of what the rules required. Intelligence investigators and prosecutors were not only allowed but required to talk to each other. While prosecutors could not direct intelligence investigations, they could provide “guidance,” and investigators were required to inform prosecutors if their investigations turned up information about serious crimes (and to provide them with monthly briefings in any event). Investigators had to get higher-level approval to share certain other kinds of information, but that’s hardly the same thing as prohibiting any communication.
The third flaw in the logic chain is the notion that “the wall” – or, more accurately, officials’ perception of a wall – led to 9/11. Carrie doesn’t make this claim; in fact, she doesn’t mention any specific national security damage that resulted from the wall. In any event, the 9/11 Commission Report makes pretty clear that the failures to communicate critical information in advance of the attack had little to do with the Justice Department’s rules and everything to do with bureaucratic incompetence.
Finally, we must briefly push back on Carrie’s statement that “changing FISA’s standard to a ‘significant purpose’ . . . has been overwhelmingly understood as an important substantive correction.” We may be traveling in somewhat different circles, but that hasn’t been our experience. Rather, among colleagues in our field, this change is widely viewed as the moment in which FISA became an existential threat to the Fourth Amendment’s warrant requirement in ordinary criminal cases. The “correction” that’s needed is to restore the primary purpose test.