On Friday the government filed its brief in opposition to the Electronic Privacy Information Center’s mandamus petition to the Supreme Court, No. 13-58, in which EPIC is challenging the legality of the FISC’s “bulk telephony metadata” orders under section 215 of the PATRIOT Act, 50 U.S.C. 1861.

Jurisdictional/Preclusion Arguments to Dismiss the Petition

The majority of the government’s brief is devoted to arguing that EPIC has not satisfied the traditional criteria for the extraordinary relief of mandamus directly from the Supreme Court.

The government’s broadest argument in this respect is that the FISA statute prohibits anyone other than the recipient of the order itself (here, Verizon) from challenging the statutory basis for the order.  (EPIC is only challenging the statutory authority for the order—it has not raised any constitutional claim.)  The government makes a similar “preclusion” argument in support of its motion to dismiss the ACLU’s statutory claim in its lawsuit filed in the Southern District of New York.  If that argument were correct (but see the ACLU’s response at pages 18-26 of its brief), then the statutory claims in both lawsuits–and in other such cases, including the First Unitarian Church case in the Northern District of California–would be dismissed.

The Supreme Court does not need to reach that broad argument in the EPIC case, however, because the government has a much stronger, narrower argument—namely, that if FISA does not categorically preclude a customer from ever suing to challenge the statutory authority for a 215 order to a telecom provider, then EPIC could sue the NSA in district court, as the ACLU and First Unitarian Church have done; and if EPIC could sue in district court, then mandamus relief directly in the Supreme Court would be inappropriate, since the general rule is that such an extraordinary remedy is unavailable unless (at a minimum) the party seeking it does not have other adequate means to attain a remedy that would provide the relief it seeks. 

In its petition, EPIC argues (p.17) that filing a district court action would be futile because section 1861(e) would “immunize” the defendants.  I don’t think this is an especially persuasive argument.  Section 1861(e) simply provides that “[a] person who, in good faith, produces tangible things under an order pursuant to this section”—that is to say, Verizon—“shall not be liable to any other person for such production.”  But even if section 1861(e) would preclude a suit by EPIC for injunctive relief against Verizon (and it’s not at all clear that it would), it certainly would not preclude an injunction prohibiting the NSA and FBI from retaining or using the metadata in question–an injunction that presumably would offer EPIC the relief it is seeking.

More to the point, even if EPIC or the federal government were correct that Congress intended to preclude telecom customers such as EPIC from obtaining effective relief in district court on a claim that a Section 215 order was contrary to statute, then presumably Congress would have intended to preclude such relief–or to confer “immunity” to the defendants–in response to a mandamus petition, as well.  To be sure, in cases where preclusion of all such relief would raise serious constitutional questions–such as where Congress endeavors to cut off constitutional claims–the Court will often construe statutes to leave available an outlet for such claims in the Supreme Court.  Cf. Felker v. Turpin.  But in a case such as this one, where Congress did not have to provide customers with any access to court at all (since EPIC’s is only a statutory claim), it would be deeply counterintuitive to conclude that Congress chose to give such customers access to the courts for effective relief only by way of mandamus directly to the Supreme Court.  (Steve tells me that it might not be so counterintuitive after all.  He might follow up with a post on this question.)

I therefore think it’s likely the Court will deny the EPIC petition on that narrower ground–based on the conclusion that either Congress intended to preclude customers from bringing statutory challenges to Section 215 orders altogether or Congress intended to permit such challenges (such as the ACLU’s) in district court under the Administrative Procedure Act.  It would not be necessary for the Court to choose between these two options, because either way mandamus would not be appropriate.

If the Court denies the EPIC petition on that narrower ground, EPIC presumably would follow the ACLU’s lead and file a suit in district court seeking relief against the NSA and/or the FBI.

Arguments on the Merits of the Statutory Question

Although the Court is likely to dismiss the petition, the government’s newly filed brief is also significant because of what the Office of the Solicitor General has to say on the merits of the central, statutory question—namely, whether the FISC was correct to find that there were “reasonable grounds to believe that the tangible things sought” from Verizon and other service providers (the metadata of all domestic phone calls) were “relevant to an authorized [FBI] investigation . . . to protect against international terrorism.”  50 U.S.C. 1861(b)(2)(A), (c)(1).

In my earlier post I offered some skeptical thoughts about most of the government’s statutory arguments, some of which appear in the EPIC brief in opposition.  I won’t repeat those counterarguments here.

I think it’s worth noting, however, that the OSG brief does not include at least two arguments that have appeared in the Department’s earlier briefs in the district court—including the argument that was at the heart of the latest FISC decision.

First, the latest government brief in the ACLU Southern District case argues that because fourteen FISC judges have concluded that the statutory “relevance” criterion is satisfied, the courts must defer to that judgment.  In the ACLU case, the government brief reasons that because the criterion is question is whether there are reasonable grounds to believe that the [records] are relevant” to FBI counter-terrorism investigations,” a federal court could only hold to the contrary if it were prepared “to conclude that the fourteen FISC judges who repeatedly issued those orders lacked any reasonable basis for doing so,” a proposition that DOJ calls “self-defeating.”

I trust that the logical flaw in this argument is self-evident.  The question at issue is whether there were reasonable grounds to believe the records subject to the metadata orders were relevant to an FBI investigation—not whether the FISA judges were “reasonable” in answering that question in the affirmative.  Nor is there any other reason for a federal court to defer to the FISC’s conclusion, no matter how many different judges affirmed it, especially since (as I explained in my previous post) it appears that the 14 judges in question might never have been presented with the best arguments against the legality of the orders.

It is not surprising, therefore, that the Solicitor General does not rely upon this argument, nor argue that the FISC is owed any deference on other grounds, either.

Second, Judge Eagan of the FISC concluded that all domestic telephony metadata is “relevant” to FBI counterterrorism investigations because obtaining all such records is necessary in order to permit NSA to do the sorts of algorithmic data searches that allow the agency to determine “connections between known and unknown international terrorist operatives.”  This same argument appears at pages 8-9 and 12-14 of the government’s White Paper, and in the government’s S.D.N.Y. briefs.

But the OSG brief does not mention this argument, or invoke any notion of “necessity”—and with good reason.  Not only is “necessity for an NSA computer data search” not the statutory standard, it is also a standard the FBI almost certainly could not satisfy.  As I wrote in my last post, “even if access to such vast reservoirs of data makes it substantially easier for NSA computers to ferret out terrorist connections–something that could hardly be denied–such collection is not in any strong sense ‘necessary’ to the NSA’s objectives:  The agency found plenty of such terrorism connections even before this program was in place; and it would be able to discover many more such connections–even if not as many–if service providers were required to retain the records, and the agency then directed the providers to search for records on a retail basis whenever the NSA provided an ‘identifier’ (such as a terrorist phone number).”

Here, too, then, I am not surprised that the Solicitor General chose not to invoke such a flawed argument.

As I read it, the government’s primary argument is instead a textual one that I did not discuss in my previous post—an argument that turns out to be even broader than Judge Eagan’s “necessity” argument.  The government asserts (at page 30 of its brief) that any records “that facilitate the government’s use of investigative tools meet the relevance standard,” even if most of those records will turn out not to have any bearing on the “subject matter”—the substance—of an FBI counterterrorism investigation.

This argument is predicated on the fact that, in contrast to the language of other “relevance” statutes, section 1861(B)(2)(A) requires reasonable grounds to believe the tangible things sought are relevant to an FBI investigation:

[U]nlike civil discovery rules, which limit discovery to matters ‘relevant to the subject matter involved in the action,’ Fed. R. Civ. P. 26(b)(1) (emphasis added in OSG brief), Section 1861 requires only that the records be relevant to an ‘authorized investigation.’  50 U.S.C. 1861(b)(2)(A).  That suggests that records that facilitate the government’s use of investigative tools meet the relevance standard.

The government’s White Paper puts the point this way:  Because Section 215 refers to records “relevant to an authorized investigation,” it covers not only information directly relevant to the authorized objects of FBI investigations (such as international terrorism), “but also information relevant to the investigative process or methods employed in reasonable furtherance of such national security investigations” (emphasis added).  And because the NSA’s possession of all domestic telephony metadata would obviously “facilitate” NSA’s data-mining to some degree, those data are relevant to the “investigative process or methods,” and thus on this view are covered under section 1861(b)(2)(A).

This textual and comparative argument is one that I think we’re likely to see more of, if and when the district court litigation proceeds on the merits.  So here are three initial, very preliminary reactions to it:

a.   Most importantly, the government’s proposed standard would appear to prove too much.  After all, wouldn’t governmental access to virtually all records, of all kinds, “facilitate the government’s use of [its] investigative tools” in a terrorism investigation?  The more information about individuals the NSA has, the more effective its diagnostic tools would be.

The government is aware of this significant problem with its argument.  Therefore, its brief tries to explain why telephony metadata are different:

[T]elecommunications records have characteristics not common to other types of records—specifically, their highly standardized and inter-connected nature—that make them readily susceptible to analysis in large datasets to bring previously unknown connections between and among individuals to light.  The same cannot be said of myriad other types of records that might be subject to a Section 1861 order.

 It’s not clear how many other types of records this distinction would exclude from the statutory coverage under the government’s theory.  Many other records are also “highly standardized and inter-connected”—particularly records about human interactions.   (See page 14 of the latest ACLU brief.)

But more fundamentally, it’s not clear why records must be “highly standardized and inter-connected” in order to be “relevant” to an FBI investigation under the government’s test.  To be sure, “highly standardized an inter-connected” records may be more useful—more helpful to make for effective data searches—than are other records.  But possession of other records, too, would “facilitate the government’s use of investigative tools” in a terrorism investigation.  If, as the government appears to argue, records are “relevant” for 1861 purposes whenever they “facilitate the government’s use of investigative tools,” then it would appear that all record sets would meet that standard, even if certain data sets are more “facilitative” than others.

Indeed, why limit it to records of individuals?  The statute authorizes an order to produce any “tangible things.”  As the ACLU rhetorically asks in response to the government’s “facilitate investigations” argument:  “Does the statute allow the government to compel a university to turn over technology that would allow more efficient or revealing analysis?”  It’s hard to imagine this is what Congress had in mind.

b.  The government does not cite any legislative history suggesting that Congress intended that a business record would be “relevant” under the statute so long as it merely “facilitate[s] the government’s use of investigative tools” (let alone facilitate such tools prior to the FBI’s receipt of any of the information).  I don’t know what the legislative history shows in that regard.  But this much does seem clear:  Congress apparently adopted the “relevant to an authorized investigation” phrasing in 1861(b)(2)(A) from several other statutes that use basically the same formulation, including:

— 12 U.S C. 5113(a)(2) (Director of Consumer Financial Protection Bureau can require mortgage loan originator to provide evidence that “may be relevant or material to an investigation of such loan originator for compliance with the requirements of this chapter”)

— 18 U.S.C. 1968(a) (“Whenever the Attorney General has reason to believe that any person or enterprise may be in possession, custody, or control of any documentary materials relevant to a racketeering investigation, he may, prior to the institution of a civil or criminal proceeding thereon, issue in writing, and cause to be served upon such person, a civil investigative demand requiring such person to produce such material for examination.”)

— 18 USC 3123(a)(1) (court shall enter an ex parte order authorizing the installation and use of a pen register or trap and trace device if it finds “that the attorney for the Government has certified to the court that the information likely to be obtained by such installation and use is relevant to an ongoing criminal investigation”). 

— 18 USC 3486(a)(1)(B)(i) (Attorney General may issue subpoena for “the production of any records or other things relevant to the investigation” of a Federal health care offense)

Most notably, Congress presumably borrowed the language of Section 215 from 18 USC 2709(b)(1), which authorizes the FBI Director to issue a “national security letter” to a telecom service provider requesting customer metadata (principally, a customer’s local and long distance toll billing records) “if the Director (or his designee) certifies in writing to the wire or electronic communication service provider to which the request is made that the . . . records sought are relevant to an authorized investigation to protect against international terrorism.”

It will be important, therefore, to determine whether Congress and/or the courts have read the “relevant to an investigation” language of such other statutes to encompass information that merely “facilitates the government’s use of investigative tools”information that is not relevant to the subject matter of the investigation in question but that is relevant (in the words of the White Paper) to the investigative process or methods employed in reasonable furtherance of that investigation.  Could the FBI use a national security letter under section 2709, for instance, to request all of a service provider’s customer’s toll records—without a FISC order–merely because having the entire set of such metadata would facilitate FBI investigations?  I would be surprised if 2709 or any of the other parallel statutes have been so construed . . . but I have not researched the question in any detail.

c.  Finally, if it were the case, as OSG argues, that a record is “relevant” for purposes of section 1861(b)(2)(A) so long as it “facilitates the government’s use of investigative tools” in a terrorism investigation, then it’s not clear that Congress’s 2006 amendment of Section 215 would have had much effect.  Between 2001 and 2006, Section 215 only required the FBI to demonstrate that the tangible things were “for an investigation to protect against international terrorism”—a very undemanding standard that the FISC orders here probably would have satisfied.  The whole point of Congress’s 2006 addition of the “relevance” standard was to meaningfully increase the government’s burden.  But if it were enough for the FISC to find that it is reasonable to believe the records would facilitate the government’s use of investigative tools,” it is hard to imagine any records a government would seek that would not be “relevant” to an FBI investigation . . . in which case the 2006 amendment would have accomplished little or nothing.

* * *

Just to be clear:  Nothing I’ve written here or in my last post speaks to the question of whether it would be a good idea — or constitutional — for Congress to authorize the NSA to obtain all such telephony metadata.  And the question I’ve written about–whether current law authorizes the FISC orders–will be rendered moot if and when Congress acts to specify whether and under what circumstances such orders can be issued.  Numerous members of Congress have offered or are drafting statutory amendments.  Enactment of any of them would presumably resolve the statutory question going forward.  But if one or more courts reach the statutory question before Congress acts, I would not be surprised if at least some of those courts rule for the plaintiffs and enjoin operation of the program.