As has often been noted, the White Paper the Bush Administration released on January 19, 2006 largely repeats the analysis Jack Goldsmith did in his May 6, 2004 OLC opinion on the warrantless wiretap program. So I decided to compare the two documents.
Not only did such a comparison help me see things in both documents I hadn’t seen before. But there are a number of things that appear in the White Paper but not the unredacted parts of the opinion. Some of this, such as Administration statements after the warrantless wiretap program was exposed in 2005, simply serve as the publicly acceptable discussion of the program. Yet in one case–the White Paper’s discussion of how the Hamdi decision affected the program–this probably repeats a discussion in another, still classified, Goldsmith opinion he wrote the day before he left on July 17, 2004. Then there’s a bunch of information that appears (in both redacted and unredacted form) in the Goldsmith opinion but not the WP. As I discuss below, I think there are a number of reasons for this.
I should warn that I did this in about a day or so, so I certainly may have misstated what’s in Goldsmith’s memo. Let me know if you catch anything like that.
Goldsmith’s memo is organized this way:
Analysis [of whether the illegal wiretap program is legal under 5 different criteria]
I. Executive Order 12333
II. Statutory Analysis (of FISA and Title III wiretap laws)
III. Completely redacted criterion*
IV. Completely redacted criterion*
V. Fourth Amendment (including extensive discussion of why the current threat makes the illegal program a reasonable search)
*If I had to guess what the two completely redacted criteria are, I’d say one is the Defense Appropriation of 2004, which prohibited data mining of US data, and one is the First Amendment.
The bolded subjects above don’t appear in the WP. The exclusion of some of this–the discussion of how the program works, for example–is dismissed in the WP by saying it cannot be discussed in an unclassified document. The EO 12333 discussion, which presumably pertains in part to the wiretapping of US persons overseas, didn’t seem to be the big public concern after the program was revealed (or maybe the WP didn’t want to admit that limits on wiretapping Americans were just pixie dusted away). And some of these subjects–such as the Defense Authorization, if my guess that it’s one of the totally redacted criteria is right–were no longer operative in 2006 when the WP was issued.
In general, Goldsmith (and the WP) replace John Yoo’s authorization of the program under Article II with what he calls “new analysis” finding that the Afghan AUMF bestowed on the President full Commander in Chief powers, which in the process meant his war powers trumped FISA. The formula isn’t much more sound than what we suspect Yoo to have said, but it gives Goldsmith lots of places to insert wiggle room into interpretations of FISA, for example, arguing that the principle of constitutional avoidance suggests that the purported conflict between the AUMF and FISA must be resolved to make sense constitutionally which, in Goldsmith’s book, means a tie goes to the Commander in Chief.
The focus on the AUMF allows both documents to rehearse a long history of wartime wiretapping that just happens to magically skip the Vietman-era wiretapping that FISA was written to prohibit.
In addition, Goldsmith (and the WP) argues that the importance of the government’s interest in wiretapping al Qaeda makes the warrantless program “reasonable” under the Fourth Amendment. Note, this is almost certainly a departure from John Yoo’s November 2, 2001 Fourth Amendment based argument, given how closely that opinion seems to cling to his October 23, 2001 Fourth Amendment evisceration opinion, and given Goldsmith’s decision not to rely on that opinion on page 100. In the Fourth Amendment discussion, Goldsmith gives very extensive (but entirely redacted) information on the threats that justify such wiretapping; the WP effectively just says “trust us.”
How They Define the Target of the AUMF
Now, in their discussions of the war on terrorism, there are two differences I noted. One is very slight–but I find very intriguing. The WP describes the people the government is permitted to wiretap this way:
The President has acknowledged that, to counter this threat, he has authorized the NSA to intercept international communications into and out of the United States of persons linked to al Qaeda or related terrorist organizations. The same day, the Attorney General elaborated and explained that in order to intercept a communication, there must be “a reasonable basis to conclude that one party to the communication is a member of al Qaeda, affiliated with al Qaeda, or a member of an organization affiliated with al Qaeda.”
That is, the WP claims the NSA can only wiretap people with known ties to al Qaeda.
But in his highly-classified memo, Goldsmith assessed the President had,
the authority to intercept the content of international communications “for which, based on the factual and practical considerations of everyday life on which reasonable and prudent persons act, there are reasonable grounds to believe … [that] a party to such communication is a group engaged in international terrorism, or activities in preparation therefor, or any agent of such a group,” as long as that group is al Qaeda, an affiliate of al Qaeda or another international terrorist group that the President has determined both (a) is in armed conflict with the United States and (b) poses a threat of hostile actions within the United States;
Obviously, Goldsmith’s version would permit the wiretapping of suspected terrorists without clear ties with Al Qaeda–perhaps a group like Hamas or Hezbollah or whatever other group the President determines in secret is in armed conflict with the US (I wonder: does dropping a drone on a group equate to them being in armed conflict with the US?). That also exceeds the terms of the Afghan AUMF (as I have pointed out, the Iraq AUMF, which in typically corrupt fashion emphasizes terrorism, talks about terrorism more generally).
And while this is probably unrelated, consider the difference in how the WP and Goldsmith conceived of the threats described in Keith (the SCOTUS wiretap case that found Nixon-era wiretapping to violate the Fourth Amendment). Goldsmith said that Keith only applied to “investigations of purely domestic threats to security–such as domestic terrorism,” whereas the White Paper says it only applied to “investigations of wholly domestic threats to security—such as domestic political violence and other crimes” [my emphasis]. Perhaps Alberto Gonzales and Steven Bradbury just edited Goldsmith’s original formulation to avoid worrying their base, which includes a great number of domestic terrorists. Perhaps Goldsmith saw the prohibition on domestic threats more broadly (and AGAG at least envisioned wiretapping people he considered terrorists with no political side–yeah, Earth First, I’m thinking of you!). Also, I think “purely” domestic threats is a somewhat broader limit on the prohibition on domestic wiretapping than “wholly” domestic threats, as it’d be pretty easy for a mental midget like AGAG to argue that Greenpeace had a partly foreign component that therefore permitted wiretapping.
Now, both the WP and Goldsmith distinguish the terrorists they’re authorizing wiretaps for from the people covered by Keith by claiming there is no First Amendment aspect to the alleged terrorists or those who speak with them. Here’s the WP version of the discussion (footnote 2 on page 9), which takes the Goldsmith discussion, also in a footnote, on page 40-41 almost verbatim.
Keith made clear that one of the significant concerns driving the Court’s conclusion in the domestic security context was the inevitable connection between perceived threats to domestic security and political dissent. As the Court explained: “Fourth Amendment protections become the more necessary when the targets of official surveillance may be those suspected of unorthodoxy in their political beliefs. The danger to political dissent is acute where the Government attempts to act under so vague a concept as the power to protect ‘domestic security.’” Keith, 407 U.S. at 314; see also id. at 320 (“Security surveillances are especially sensitive because of the inherent vagueness of the domestic security concept, the necessarily broad and continuing nature of intelligence gathering, and the temptation to utilize such surveillances to oversee political dissent.”). Surveillance of domestic groups raises a First Amendment concern that generally is not present when the subjects of the surveillance are foreign powers or their agents.
But where the WP version ends, Goldsmith’s discussion continues in maybe 4-6 lines of redacted discussion, then ends with this:
One of the important factors driving the Supreme Court’s conclusion that the warrant requirement should apply in the domestic security context is thus simply absent in the foreign intelligence realm.
Frankly, pretty much any of the plaintiffs in the newly-reinstated ACLU suit, not to mention the increasing number of young Muslim men who get entrapped after their speech suggests they’re a threat, would be able to challenge this assertion to show that the program does raise significant First Amendment issues, not least given the difference in the way fundamentalist Muslim terrorists are treated from how fundamentalist Christian terrorists (with almost the sole exception of the Hutaree) are treated.
But Are We Even at War?
There are two curious divergences in the two papers’ discussion of whether we’re even at war. In addition to Keith, both look to a case called Barreme, which pertained to whether or not the President could order the military to seize ships during a “Quasi War” with France in 1799 (a close equivalent might be a suit on whether the President can allow Halliburton to do business in defiance of sanctions against Iran). Goldsmith’s discussion of the case is far more extensive than what appears in the WP. And as part of it, he claimed that Barreme was distinguished because it wasn’t entirely clear to SCOTUS that the US and France were really, fully at war.
The Court’s decision was fundamentally based on the premise that the state of affairs with France was not sufficiently akin to a full-scale war for the President to invoke under his own inherent authority the full rights of war that, in other cases, he might have at his disposal. As a result, he required the special authorization of Congress to act.
It’s not the way it worked out, but it is conceivable to imagine a world post-9/11 in which Republicans challenged a Democratic President as to whether we were really, fully at war against a band of terrorists. In fact, I think they did after the 1998 Embassy bombings. (Of course, a Democrat wouldn’t have ignored warnings about “Osama bin Laden determined to strike in the US.”)
Meanwhile, written post-Hamdi, the WP has no such doubts. But it does introduce this bizarre argument, not present in the unredacted Goldsmith, that says the AUMF grants even greater power to the Commander in Chief than an actual, old-fashioned, declared war does.
The contrary interpretation of section 111 [giving the President a 15 day window after the declaration of war to wiretap outside of FISA] also ignores the important differences between a formal declaration of war and a resolution such as the AUMF. As a historical matter, a formal declaration of war was no longer than a sentence, and thus Congress would not expect a declaration of war to outline the extent to which Congress authorized the President to engage in various incidents of waging war. Authorizations for the use of military force, by contrast, are typically more detailed and are made for the specific purpose of reciting the manner in which Congress has authorized the President to act.
I will come back to the more general discussion of the 15-day window–I’ll show why, having been caught breaking the law, the Bushies may have wanted to pretend there was no possible 15-day period in this case. But for the moment, consider how, between Goldsmith and the WP, they’re both trying to claim all the powers of the Commander in Chief during war, but maybe not always.
The Use of SIGINT Under War
One of the most interesting things I first realized by comparing these two documents is how they both describe the goal of the wiretap program
The WP latches onto the AUMF’s unwise authorization of the President to “determine” who hit us on 9/11 and with it emphasizes that that language permits the President to use wiretapping to identify and locate the enemy.
The terms of the AUMF not only authorized the President to “use all necessary and appropriate force” against those responsible for the September 11th attacks; it also authorized the President to “determine” the persons or groups responsible for those attacks and to take all actions necessary to prevent further attacks. AUMF § 2(a) (“the President is authorized to use all necessary and appropriate force against those nations, organizations, or persons he determines planned, authorized, committed, or aided the terrorist attacks that occurred on September 11th, 2001, or harbored such organizations or persons”) (emphasis added). Of vital importance to the use of force against the enemy is locating the enemy and identifying its plans of attack. And of vital importance to identifying the enemy and detecting possible future plots was the authority to intercept communications to or from the United States of persons with links to al Qaeda or related terrorist organizations. Given that the agents who carried out the initial attacks resided in the United States and had successfully blended into American society and disguised their identities and intentions until they were ready to strike, the necessity of using the most effective intelligence gathering tools against such an enemy, including electronic surveillance, was patent. [my emphasis]
Later, the WP uses the verbs “identify” and “pinpoint.”
The use of signals intelligence to identify and pinpoint the enemy is a traditional component of wartime military operations—or, to use the terminology of Hamdi, a “fundamental and accepted . . . incident to war,” 542 U.S. at 518 (plurality opinion)—employed to defeat the enemy and to prevent enemy attacks in the United States. [my emphasis]
I find this notable because, in spite of the fact the WP seemingly focuses exclusively on the interception of the content of communications, the verbs it uses just as easily apply to data mining and geo-location as to content. I wondered to myself whether the WP was trying to sneak in claimed authorization to do data mining, even while Bush claimed only to be talking about the collection of actual content.
And once I saw that in the WP, I realized Goldsmith uses the same language.
[Redacted name of program] is a highly classified and strictly compartmented program of electronic surveillance within the United States that President Bush directed the Department of Defense to undertake on October 4, 2001 in response to the attacks of September 11, 2001. Specifically, the program is designed to counter the threat of further terrorist attacks on the territorial United States by detecting communications that will disclose terrorist operatives, terrorist plans, or other information that can enable the disruption of such attacks, particularly the identity of al Qaeda operatives with the United States. [my emphasis]
Goldsmith used this same language on page 62 in his discussion of Youngstown.
First, the exercise of executive authority here is not several steps removed from the actual conduct of a military campaign. To the contrary, [redacted longer name of program] is an intelligence operation undertaken by the Department of Defense specifically to detect and disrupt planned attacks, largely by detecting enemy agents already within the United States. Al Qaeda has already demonstrated an ability, both on September 11 and subsequently (in such cases as Jose Padilla and Ali al-Marri) to insert agents into the United States. As explained above, the efforts [redacted shorter name of program] to intercept communications that would lead to the discovery of more such agents or other planned attacks on the United States are a core exercise of Commander-in-Chief authority in the midst of an armed conflict. [my emphasis]
Now, this reference–with its mention of an earlier discussion–is of particular interest. If I’m not mistaken, the earlier reference is redacted. In other words, it suggests that Goldsmith discussed the aspects of the program that are specifically focused on detecting additional al Qaeda figures in one of the sections that is redacted. That may be because the entire discussion about data mining in his memo is redacted, or it may be for another reason.
In any case, I find it interesting that the WP adopts language of detection without engaging in any discussion of those aspects of the program.
The Description of Congress’ Intent in FISA
While the papers’ discussion of the passage of FISA legislation could be a post of its own, there are two main differences between the papers’ treatment of the legislative record on it. In his classified opinion, Goldsmith mines every hesitation on the part of Congress and especially Attorney General Levi and Deputy Attorney General Laurence Silberman–effectively letting the executive branch stand in for the legislative history for Congress. (Though he also invokes Teddy Kennedy’s claimed doubts.) None of that shows up in the WP, suggesting that the Bush Administration was more willing to rely on partisan hacks like Silberman when they thought no one would ever see it, than they were willing to do in public. (That’s sort of odd given that Steven Bradbury used to love relying on William Rehquist’s hackish OLC opinions.)
And if I’m not mistaken, the unredacted parts of the classified memo don’t include this pretty damning admission.
That [FISA] report includes the extraordinary acknowledgment that “[t]he conferees agree that the establishment by this act of exclusive means by which the President may conduct electronic surveillance does not foreclose a different decision by the Supreme Court.” H.R. Conf. Rep. No. 95-1720, at 35, reprinted in 1978 U.S.C.C.A.N. 4048, 4064. But, invoking Justice Jackson’s concurrence in the Steel Seizure case, the Conference Report explained that Congress intended in FISA to exert whatever power Congress constitutionally had over the subject matter to restrict foreign intelligence surveillance and to leave the President solely with whatever inherent constitutional authority he might be able to invoke against Congress’s express wishes. Id. The Report thus explains that “[t]he intent of the conferees is to apply the standard set forth in Justice Jackson’s concurring opinion in the Steel Seizure Case: ‘When a President takes measures incompatible with the express or implied will of Congress, his power is at the lowest ebb, for then he can rely only upon his own constitutional power minus any constitutional power of Congress over the matter.’” Id. (quoting Youngstown Sheet & Tube Co. v. Sawyer, 343 U.S. 579, 637 (1952) (Jackson, J., concurring));
Now, it’s possible this passage is in the redacted parts of Goldsmith’s memos (though there doesn’t appear to be a logically placed redaction for it). It’s also possible that David Kris, whom we know was criticizing drafts of the WP and who cited precisely this passage in his shredding of the WP after it was published, shamed DOJ into including it in the WP. But Goldsmith’s apparent (though not definite) choice to ignore that passage doesn’t say much for his good faith when writing the memo.
And then there is a slightly different treatment of the exclusivity clause, which says that FISA really and truly is the only way the President can collect foreign intelligence. Here’s how Kris described what DOJ did in the white paper to dismiss exclusivity.
[T]he NSA surveillance program accords with the exclusivity provision because FISA “permits an exception” to its own procedures where surveillance is “authorized by another statute, even if the other authorizing statute does not specifically amend” the exclusivity provision;
Here’s how that kind of sophism looks in practice.
To be sure, the scope of this exception is rendered less clear by the conforming amendments that FISA made to chapter 119 of title 18—the portion of the criminal code that provides the mechanism for obtaining wiretaps for law enforcement purposes. Before FISA was enacted, chapter 119 made it a criminal offense for any person to intercept a communication
except as specifically provided in that chapter. See 18 U.S.C. § 2511(1)(a), (4)(a). Section 201(b) of FISA amended that chapter to provide an exception from criminal liability for activities conducted pursuant to FISA. Specifically, FISA added 18 U.S.C. § 2511(2)(e), which provides that it is not unlawful for “an officer, employee, or agent of the United States . . . to conduct electronic surveillance, as defined in section 101 of the Foreign Intelligence Surveillance Act of 1978, as authorized by that Act.” Id. § 2511(2)(e). Similarly, section 201(b) of FISA amended chapter 119 to provide that “procedures in this chapter [or chapter 121 (addressing access to stored wire and electronic communications and customer records)] and the Foreign Intelligence Surveillance Act of 1978 shall be the exclusive means by which electronic surveillance, as defined in section 101 of such Act, and the interception of domestic wire, oral, and electronic communications may be conducted.” Id. § 2511(2)(f) (West Supp. 2005).7
The amendments that section 201(b) of FISA made to title 18 are fully consistent, however, with the conclusion that FISA contemplates that a subsequent statute could authorize electronic surveillance outside FISA’s express procedural requirements. Section 2511(2)(e) of title 18, which provides that it is “not unlawful” for an officer of the United States to conduct electronic surveillance “as authorized by” FISA, is best understood as a safe-harbor provision. Because of section 109, the protection offered by section 2511(2)(e) for surveillance “authorized by” FISA extends to surveillance that is authorized by any other statute and therefore excepted from the prohibition of section 109. In any event, the purpose of section 2511(2)(e) is merely to make explicit what would already have been implicit—that those authorized by statute to engage in particular surveillance do not act unlawfully when they conduct such surveillance.
It’s not clear whether Goldsmith engages in the same language games. Goldsmith’s discussion includes the following passage (starting on 20):
On their face, [50 USC 1809 and 18 USC 25110 make FISA, and the authorization process it requires, the exclusive lawful means for the Executive to engage in "electronic surveillance," as defined in the Act for foreign intelligence purposes. Indeed, this exclusivity is expressly emphasized in section 2511(2)(f), which states that "procedures in this chapter or chapter 121 [addressing access to stored wire and electronic communications and customer records] and the Foreign Intelligence Surveillance Act of 1978 shall be the exclusive means by which electronic surveillance, as defined in section 101 may be conducted.[my emphasis; note, too, this is one of the paragraphs of the opinion that are clearly misclassified]
As I’ll ultimately show, this entire paragraph appears to be overclassified, suggesting the White House was trying to hide the admission that there is such a thing as an exclusivity provision. Unless I missed something, though, Goldsmith doesn’t appear to have engaged in the sophism above in the unredacted opinion. But, there’s a long redacted paragraph right after this discussion with a footnote, so it may be that DOJ just decided to hide his version of this neat trick.
In perhaps a related issue, the WP has this discussion on 18 USC 2511(2)(a)(ii) while the unredacted parts of Goldsmith don’t.
In addition, section 2511(2)(a)(ii) authorizes telecommunications providers to assist officers of the Government engaged in electronic surveillance when the Attorney General certifies that “no warrant or court order is required by law [and] that all statutory requirements have been met.” 18 U.S.C. § 2511(2)(a)(ii).9 If the Attorney General can certify, in good faith, that the requirements of a subsequent statute authorizing electronic surveillance are met, service providers are affirmatively and expressly authorized to assist the Government.
Given that the requests to the telecoms (with the exception of the March 11, 2004 one) reportedly took the form of a request under this statute–an assertion by the Attorney General that no warrant was required–I suspect that it was invoked more generally and therefore suspect it appears in one of the redacted passages of Goldsmith.
Odds and Ends
Finally, there are four details that suggest certain things about the larger program and/or the government’s beliefs about their forever everywhere war on terror. For example, Goldsmith’s discussion of Truong, another wiretap case, is more extensive than the WP. As part of it, Goldsmith includes the quote that getting warrants in a foreign case “would potentially jeopardize security by increasing ‘the chance of leaks and secrecy.’” Now, a purported claim to be worried about leaks was at the heart of the Bush Administration’s refusal to fulfill requirements on briefing Congress (I’ll discuss this more in my post on the 15 day exception). But for some reason, the government didn’t want to make this claim in a document that Congress would actually get to read.
Then in a discussion of Youngstown, Goldsmith quotes the line “[e]ven though ‘theater of war’ [may] be an expanding concept.” Again, we know the government (the Bush and Obama Administrations) like to claim the entire world is their unlimited power oyster. But it’s telling that the WP didn’t feel the need to repeat this to the Congress that it had just been caught flouting for four years.
Finally, in the WP there are two references to technical issues too classified to discuss. First, there’s a discussion of how the need for speed and agility required that Bush blow off FISA.
The second serious constitutional question is whether the particular restrictions imposed by FISA would impermissibly hamper the President’s exercise of his constitutionally assigned duties as Commander in Chief. The President has determined that the speed and agility required to carry out the NSA activities successfully could not have been achieved under FISA.18 Because the President also has determined that the NSA activities are necessary to the defense of the United States from a subsequent terrorist attack in the armed conflict with al Qaeda, FISA would impermissibly interfere with the President’s most solemn constitutional obligation—to defend the United States against foreign attack.
18 In order to avoid further compromising vital national security activities, a full explanation of the basis for the President’s determination cannot be given in an unclassified document.
And then, there’s a discussion of the ways in which technical changes made FISA outdated in ways that Congress couldn’t really fix.
Third, certain technological changes have rendered FISA still more problematic. As discussed above, when FISA was enacted in 1978, Congress expressly declined to regulate through FISA certain signals intelligence activities conducted by the NSA. See supra, at pp. 18-19 & n.6.20 These same factors weigh heavily in favor of concluding that FISA would be unconstitutional as applied to the current conflict if the canon of constitutional avoidance could not be used to head off a collision between the Branches.
20 Since FISA’s enactment in 1978, the means of transmitting communications has undergone extensive transformation. In particular, many communications that would have been carried by wire are now transmitted through the air, and many communications that would have been carried by radio signals (including by satellite transmissions) are now transmitted by fiber optic cables. It is such technological advancements that have broadened FISA’s reach, not any particularized congressional judgment that the NSA’s traditional activities in intercepting such international communications should be subject to FISA’s procedures. A full explanation of these technological changes would require a discussion of classified information.
Neither of these topics appears in the unredacted parts of Goldsmith’s memo. But both are assuredly there in some form. Both technological changes and quickness are likely to appear in the entirely redacted sections after page 70 or the entirely redacted pages from 75-99.