Statement of Associate Deputy Attorney General David S. Kris
Before the Senate Judiciary Committee
Concerning the Foreign Intelligence Surveillance Act
Presented on September 10, 2002
Mr. Chairman, Senator Hatch, and Members of the Committee: Thank you for the opportunity to testify about the governments first appeal to the Foreign Intelligence Surveillance Court of Review. Obviously, the record in the appeal is classified, because the underlying FISA applications are classified. But I am pleased to provide as much information as possible about the appeal in this open hearing. The May 17 opinion of the Foreign Intelligence Surveillance Court (the FISC), as well as a redacted version of the governments brief on appeal, are not classified, and have already been provided to the Committee. Those two documents set forth the main legal arguments pro and con, albeit without reference to the facts of any particular case.
At the request of your staff, I have focused on three specific questions: First, the question of what is at stake in this appeal, and the differences between the Intelligence Sharing Procedures proposed by the government and those approved by the FISC in its May 17 order. Second, the FISCs legal reasoning as well as our interpretation of FISA and the USA Patriot Act; here I will also discuss the practical implications of our legal reasoning. Third and finally, concerns about the accuracy of FISA applications which are raised in the FISCs May 17 opinion. At the Committees request, I have prepared specifically to address those three issues. I know that there are many other FISA-related issues in the air today, but I must say that I have not specially prepared to address them. The appeal is more than enough to tackle in one sitting.
To frame the issues properly, I have to review some background about FISA, and describe what has - and has not - changed as a result of the USA Patriot Act. Here is what has not changed. As always, FISA continues to require advance judicial approval for almost all electronic surveillance and physical searches. As always, every FISA application must be personally signed and certified by a high-ranking and politically accountable Executive Branch official, such as the Director of the FBI. As always, every FISA application must also be personally signed and approved by the Attorney General or the Deputy Attorney General. As always, a FISA target must be an "agent of a foreign power" as defined by the statute - a standard that for United States persons requires not only a connection to a foreign power, but also probable cause that the person is engaged in clandestine intelligence activities, international terrorism, sabotage, or related activity.
Let me use terrorism as an example of this last point. A United States person - a citizen or green card holder - cannot be an "agent of a foreign power" under the rubric of terrorism, and therefore cannot be a FISA target, unless the government shows, and the court finds, probable cause that he is "knowingly engaged in" or "prepar[ing]" to engage in "international terrorism" for or on behalf of a foreign power. 50 U.S.C. § 1801(b)(2)(C).
FISA defines "international terrorism" to require, among other things, the commission of "violent" or "dangerous" acts that either "are a violation" of U.S. criminal law, or that "would be a criminal violation" if they were committed here. 50 U.S.C. § 1801(c). Flying an airplane into the World Trade Center is a crime under U.S. law, and therefore could qualify as an international terrorist act. Flying an airplane into the Eiffel Tower may not be a crime under U.S. law, but it could still qualify as international terrorism because it would be criminal if the Eiffel Tower were within our jurisdiction. To search or surveil a U.S. person under the terrorism provision of FISA, we have to show that he is "knowingly engaged" in committing, or "prepar[ing]" to commit, such a criminal act. And, of course, FISA imposes other requirements that are not required in an ordinary criminal case.
None of this was changed by the USA Patriot Act. The USA Patriot Act did certainly change the allowable "purpose" of a search or surveillance - the reasons "why" FISA may be used. But it is also accurate to say - though perhaps in need of elaboration - that the Act did not fundamentally change the "who," the "what," the "where," the "when" or the "how" of FISA surveillance.
What is at Stake in the Appeal
The first issue you identified for me concerns what is at stake in the appeal. What is at stake is nothing less than our ability to protect this country from foreign spies and terrorists. When we identify a spy or a terrorist, we have to pursue a coordinated, integrated, coherent response. We need all of our best people, intelligence and law enforcement alike, working together to neutralize the threat. In some cases, the best protection is prosecution - like the recent prosecution of Robert Hanssen for espionage. In other cases, prosecution is a bad idea, and another method - such as recruitment - is called for. Sometimes you need to use both methods. But we cant make a rational decision until everyone is allowed to sit down together and brainstorm about what to do. That is what we are seeking.
Let me draw a medical analogy. When someone has cancer, sometimes the best solution is surgery to cut the tumor out. Other times, its chemotherapy. And in some cases you need both. But who would go to a hospital where the doctors cant sit down and talk to each other about whats best for the patient? Thats bad medicine. And that is what were trying to change.
Now let me describe the more technical aspects of what is at stake. The Intelligence Sharing Procedures proposed by the Department in March 2002 would have permitted a full range of coordination between intelligence and law enforcement officials, including both (1) information-sharing and (2) advice-giving. The FISC accepted in full the Departments standards governing information-sharing. Under those standards, the FBI must keep prosecutors informed of "all information" that is "necessary to the ability of the United States to investigate or protect against foreign attack, sabotage, terrorism, and clandestine intelligence activities." Thus, absent special limits imposed in particular cases, intelligence officials may share a full range of information with their law enforcement counterparts, including federal prosecutors.
Sharing information from intelligence to law enforcement, however, is only half of the equation. The other half is advice about the conduct of the investigation going back the other way, from law enforcement to intelligence officials. For example, if intelligence agents inform a prosecutor that a FISA target is engaging in espionage, the prosecutor may want to advise the intelligence agents to obtain the targets financial records - for example, to determine whether he has been receiving cash deposits from a hostile foreign government. In its May 17 opinion, the FISC limited the amount of advice that can be given. While the precise extent of those limits remain somewhat opaque, even after the filing of a motion for clarification, the FISC clearly did not give us the authority that we think is appropriate under the law.
Finally, and perhaps most importantly, the FISC imposed a "chaperone" requirement, which would prohibit intelligence agents from consulting with prosecutors unless they first invite the Departments Office of Intelligence Policy and Review (OIPR) to participate. For its part, OIPR must attend the consultation unless it is "unable" to do so. That is an enormous impediment, especially because OIPR is located in Washington and the cases arise all over the country. Investigations of foreign spies and terrorists are - or at least should be - dynamic and fast-paced. Agents and lawyers need to meet and talk on the phone 5, 10, or 20 times a day, day after day, to move the investigation forward. To illustrate with an example, let me return to the medical analogy. The problem with the FISCs order is that it does not recognize that surgery, as much as chemotherapy, can be used to treat cancer - that the surgeon, as much as the oncologist, is trying to save the patient. The order says that before the oncologist can even talk to the surgeon, he has to call the hospital administrator and invite him to attend the consultation - even if the doctors and the patient are located in Los Angeles, and the hospital administrator is located in Washington. That is obviously an unworkable system.
The next question is why, as a matter of law, we disagree strongly, but respectfully with the Foreign Intelligence Surveillance Court (FISC)s decision and legal analysis in this matter. Our legal arguments are laid out fully in our brief, but will be summarized here. First, let me provide a legal framework. Since its enactment, FISA has required that some part of the governments purpose for conducting surveillance - whether it be the purpose, the primary purpose, or a significant purpose - must be to obtain what is called "foreign intelligence information." That raises two questions: First, what is "foreign intelligence information?"; and second, how much of a foreign intelligence purpose is required?
With respect to the first question, courts have generally (not always) indicated that a "foreign intelligence" purpose is a purpose to protect against foreign threats to national security, such as espionage and terrorism, using methods other than law enforcement. In other words, they have drawn a distinction between a foreign intelligence purpose and a law enforcement purpose. (This is a distinction that we think is false, as I will explain shortly.) In keeping with that approach, the courts - including the FISC - have generally evaluated the governments purpose for using FISA by evaluating the nature and extent of coordination between intelligence and law enforcement officials. The more coordination that occurs, the more likely courts are to find a law enforcement purpose rather than a foreign intelligence purpose. Under pre-Patriot Act law, if the law enforcement purpose became primary, the surveillance had to stop.
We have two arguments on appeal that correspond to these two questions. The first concerns the definition of "foreign intelligence information," which under FISA includes information needed to "protect" against espionage and international terrorism. We maintain that one way to achieve that protection is to prosecute the spies and terrorists and put them in jail - in other words, that surgery, as much as chemotherapy, can help treat cancer. Prosecution is not the only way, but its one way to protect the country. As a result, information needed as evidence in such a prosecution is itself "foreign intelligence information" as defined by the statute. In support of that argument, we rely on the original language of FISA and also on Section 504 of the USA Patriot Act, which created 50 U.S.C. §§ 1806(k) and 1825(k). As the Chairman of this Committee stated in describing Section 504:
In addition, I proposed and the Administration agreed to an additional provision in Section 505 [later changed to Section 504] that clarifies the boundaries for consultation and coordination between officials who conduct FISA search and surveillance and Federal law enforcement officials including prosecutors. Such consultation and coordination is authorized for the enforcement of laws that protect against international terrorism, clandestine intelligence activities of foreign agents, and other grave foreign threats to the nation. Protection against these foreign-based threats by any lawful means is within the scope of the definition of foreign intelligence information, and the use of FISA to gather evidence for the enforcement of these laws was contemplated in the enactment of FISA. The Justice Department's opinion cites relevant legislative history from the Senate Intelligence Committee's report in 1978, and there is comparable language in the House report.
147 Cong. Rec. S11004 (Oct. 25, 2001) (emphasis added) (statement of Senator Leahy). This same argument - that "foreign intelligence information" includes information sought for use in prosecutions designed to protect against foreign spies and terrorists - was repeated by the Chairman and Members of the Committee in the publicly available letter they sent to the FISC in July of this year:
We appreciate that "foreign intelligence information" sought under FISA may be evidence of a crime that will be used for law enforcement purposes to protect against international terrorism, sabotage, and clandestine intelligence activities by foreign powers. * * * * [Quoting the 1978 House Report, the letter states that FISA] "explicitly recognizes that information which is evidence of crimes involving clandestine intelligence activities, sabotage, and international terrorism can be sought, retained, and used pursuant to this bill." * * * * Coordination between FBI Agents and prosecutors is essential to ensure that the information sought and obtained under FISA contributed most effectively to protecting the national security against such threats.
Letter of July 31, 2002 (available at 2002 WL 1949260). This corresponds exactly to the governments principal argument on appeal.
Our second argument concerns the "significant purpose" amendment. As I mentioned, under prior law, foreign intelligence had to be the primary purpose for a FISA, and "foreign intelligence" was understood to exclude information needed for law enforcement. Thus, law enforcement could be a "significant" purpose for using FISA, but it could not be the "primary purpose." Coordination between intelligence and law enforcement personnel had to be restrained in keeping with that limit. The Patriot Act changed the old law, allowing the intelligence purpose to drop from primary to significant, and correspondingly allowing the law enforcement purpose to rise from significant to primary. What that means, at ground level, is that more coordination between intelligence and law enforcement officials should be tolerated, because even if a court would find that extensive coordination made law enforcement the primary purpose, the surveillance is still lawful under FISA. I am not sure we will have many cases in which our primary purpose is law enforcement, but the important thing is that even if we do, or the courts find that we do, the surveillance will not be at risk.
The Congressional record is replete with statements acknowledging that Members understood the implications of this change to "significant" purpose. See, 147 Cong. Rec. S10593 (Oct. 11, 2001) (statement of Senators Leahy and Cantwell); 147 Cong. Rec. S11021 (Oct. 25, 2001)(statement of Senator Feingold); 147 Cong. Rec. S11025 (Oct. 25, 2001)(statement of Senator Wellstone); Hearing of September 24, 2001, available at 2001 WL 1147486 (statement of Senator Edwards).
Indeed, it is not surprising that these Members of Congress understood the point, because the Department itself clearly described the implications of the "significant purpose" amendment in written submissions. In a letter supporting the "significant purpose" amendment sent on October 1, 2001 to the Chairs and Ranking Members of the House and Senate Judiciary and Intelligence Committees, the Department stated that the amendment would recognize that "the courts should not deny [the President] the authority to conduct intelligence searches even when the national security purpose is secondary to criminal prosecution." Letter from Assistant Attorney General Dan Bryant to the Chairs and Ranking Members of the House and Senate Judiciary and Intelligence Committees, October 1, 2001 (page 13).
Finally, outside media observers were aware of, and reported, the implications of the "significant purpose" amendment while the Patriot Act was under consideration. As Congressional Quarterly reported separately on October 8, 9, and 23, 2001: "Under the measure, for example, law enforcement could carry out a FISA operation even if the primary purpose was a criminal investigation." Congressional Quarterly, House Action Reports, Fact Sheet No. 107-33 (Oct. 9, 2001), at page 3; see Congressional Quarterly, House Action Reports, Legislative Week (Oct. 23, 2001), at page 3; Congressional Quarterly, House Action Reports, Legislative Week (October 8, 2001), at page13.
Ultimately, the courts will decide whether or not the governments legal arguments are persuasive. Those who claim that Congress never envisioned those legal arguments, however, face a steep uphill battle in light of the historical record.
Implications of the Governments Legal Arguments
Having outlined our basic legal arguments, and their support in the Patriot Act, let me make a few observation about the effect those arguments will have at ground level. Of course, predictions are limited because we have not yet been able to implement the Patriot Acts "purpose" and "coordination" amendments.
As I mentioned, under pre-Patriot Act law, the courts treated "foreign intelligence information" as if it excluded information sought for use in a criminal prosecution, apparently even if the prosecution was itself designed to protect national security against foreign threats - e.g., the prosecution of Robert Hanssen for espionage. That does not mean, however, that the government did not monitor persons such as Robert Hanssen. On the contrary, we did monitor them, albeit not for the primary purpose of prosecuting them. Even if our purpose is allowed to change, however, the scope of the surveillance will not change. As the FISC recognized in its May 17 opinion, the information sought by law enforcement officials for the prosecution of a spy or terrorist is essentially the same as the information sought by intelligence agents for a traditional counterintelligence investigation under FISA. See FISC May 17, 2002 opinion at 10 (second-to-last bullet point), 25 (first bullet point). Congress also understood that congruence when it enacted FISA in 1978. See House Report 49, 62 ("evidence of certain crimes like espionage would itself constitute foreign intelligence information as defined"). A good criminal-espionage investigation requires essentially the same information as a good counterintelligence-espionage investigation. Thus, under the Patriot Act, our purpose may change, but we will still be seeking and collecting the same information as before.*
We also claim, as a result of the "significant purpose" amendment, that FISA may be used primarily for a law enforcement purpose of any sort, as long as a significant foreign intelligence purpose remains. Considered in context, with the rest of FISAs provisions, judicial approval of this argument will also not radically change the scope of surveillance. There are several reasons why this is true:
First, of course, the "significant purpose" amendment will not and cannot change who the government may monitor. Domestic criminals - e.g., corrupt Enron executives, Bonnie and Clyde, Sammy "the Bull" Gravano, and Timothy McVeigh - cannot be FISA targets because they are not agents of foreign powers as defined by 50 U.S.C. § 1801(b). Regardless of the governments purpose, these targets simply do not satisfy FISAs probable cause requirements because they are not agents of a foreign power.
Indeed, such persons are immune from FISA for another reason: There would be no foreign intelligence purpose for monitoring them, and FISA still requires a "significant" foreign intelligence purpose. Thus, both before and after the Patriot Act, FISA can be used only against foreign powers and their agents, and only where there is at least a significant foreign intelligence purpose for the surveillance. Let me repeat for emphasis: We cannot monitor anyone today whom we could not have monitored at this time last year.
That means we are considering only a very small subset of cases in which the FISA target is an agent of a foreign power - e.g., is knowingly engaged in international terrorism on behalf of an international terrorist group - and is also committing a serious but wholly unrelated crime. In other words, we are talking about international terrorists who also engage in insider trading to line their own pockets (not to finance their terrorism), or spies who also market child pornography. I do not say that such cases could never arise; I do say that they do not arise very often. Especially in the case of U.S. persons, most agents of foreign powers are too busy carrying out their foreign intelligence missions to find time to dabble in serious but unrelated crime.
It is important to note, however, that even where such persons exist, we have always been allowed to monitor them, and we have always been allowed to share with prosecutors evidence of their unrelated crimes. See 50 U.S.C. § 1801(h)(3). Indeed, it is ironic that, because of the way courts have interpreted the term "foreign intelligence information," the governments right to share information concerning an unrelated crime was more clearly established under 50 U.S.C. § 1801(h)(3) than was the governments right to share information concerning terrorism and espionage offenses under 50 U.S.C. § 1801(h)(1). See House Report 62 (making clear that the right to share evidence of a crime with prosecutors under 50 U.S.C. § 1801(h)(3) applies only to crimes that are unrelated to the targets foreign intelligence activities). In such cases, therefore, the only real requirement changed by the Patriot Act was the one that prevented prosecutors from giving advice designed to enhance the possibility of a prosecution for the unrelated crime. The USA Patriot Act allows prosecutors to give more advice in such cases, but (with the caveat about predictions noted above) its effect on the scope of surveillance should be - at most - quite modest. Thus, while not changing who will be subject to FISA surveillance, the "significant purpose" amendment does provide a substantial benefit by allowing prosecutors and intelligence investigators to share information and advice to best coordinate their overall efforts.
Finally, there is the question of accuracy. The FISCs May 17 opinion describes two sets of FISA cases in which accuracy problems arose. I cannot discuss specifics, but I can say that the two sets of cases were unrelated. In response to these errors, the Department adopted both a short-term and a long-term response.
In the short term, we began by immediately correcting the mistakes with the court. Indeed, the FISC learned of the errors only because we brought them to its attention. We also advised the Congressional Intelligence Committees, consistent with our statutory obligation to keep them "fully inform[ed]" about the use of FISA. 50 U.S.C. § 1808(a). Finally, as the FISCs opinion reveals, the Departments Office of Professional Responsibility (OPR) opened an investigation, which is still pending, and in accord with Department policy, I will not comment on it.
For the long term, we tried to understand the structural reasons that led to inaccuracies. Here is what we discovered: The main challenge to accuracy in FISA applications is that the FBI agent who signs the affidavit describing the investigation for the court is not the agent who actually conducts the investigation. That is true because of the nature of counterintelligence investigations and FISA. By definition, every FISA investigation - an investigation in which FISA is used - is both national and international in scope, involving hostile foreign powers that target this country as a whole. As a result, any given FISA target may be part of an investigation that takes place simultaneously in, for example, New York, Los Angeles, Boston, and Houston.
Although the investigations take place all over the country, FISA applications are prepared, certified, approved, and filed with the FISC exclusively in Washington, D.C. As a result, the person who signs the FISA declaration and swears to it in the FISC is an FBI headquarters supervisor, who coordinates, but does not conduct, the field investigations he is describing for the court in his affidavit. And that is where inaccuracy can creep in: If the headquarters agent has a miscommunication with the agents in the field, his affidavit will be inaccurate.
Given that diagnosis of the problem, the solution followed logically: Require better communication between headquarters and the field. On April 5, 2001, the FBI adopted new procedures, referred to now as the "Woods Procedures," to accomplish that. The Woods Procedures are long and complex, but their basic requirement is for field agents to review and approve the accuracy of FISA applications that describe investigations occurring in their offices. In the same spirit, the Attorney General issued a memorandum in May 2001 requiring direct contact between DOJ attorneys and FBI field agents, and imposing certain other reforms as well. Both the Woods Procedures and the May 18 memo are unclassified and have previously been provided to the Committee.
I am pleased to say that the accuracy reforms have brought improved results. Perhaps the best unclassified evidence for that is a public speech given by the Presiding Judge of the FISC in April of this year, in which he said, among other things, that "we consistently find the [FISA] applications well-scrubbed by the Attorney General and his staff before they are presented to us," and that "the process is working. It is working in part because the Attorney General is conscientiously doing his job, as is his staff." It was particularly gratifying to hear the judge compliment the FBI. He said: "I am personally proud to be a part of this process, and to be witness to the dedicated and conscientious work of the FBI, NSA, CIA, and Justice Department officials and agents who are doing a truly outstanding job for all of us."
As I said, these are complex issues, and I will be happy now to answer your questions. Thank you.