The Case Against the Graham Bill

Looming in the background of the ongoing conversation over the appropriate way forward on detention (both at Guantánamo and elsewhere) is S. 3707 (the “Terrorist Detention Review Reform Act,” better known as the Graham bill), introduced last month and framed as one of the most comprehensive attempts yet at legislative resolution (or at least clarification) of many of the recurring issues in the post-Boumediene habeas litigation. 

Those who defend the Graham bill (and, more generally, the need for some legislation), tend to rely on three related arguments: (1) that the substantive, procedural, and evidentiary issues raised in detainee habeas cases are too important to be left to the “common-law process” unfolding in the D.C. courts; (2) that, to date, the D.C. courts have answered the questions that have arisen in a manner that is either inconsistent or simply incorrect; and (3) that, as such, legislation will not just stabilize the current litigation, but will also have a significant settling effect on forward-looking policy, since the relevant government actors will have a clearer sense of the substantive and evidentiary standards to which their decisions will eventually be subjected. Indeed, one can see threads of each of these arguments in the series of posts (e.g., this one) by Ben Wittes on the Graham bill (on which he consulted) over at “Lawfare.”

In the post that follows, I aim to identify a number of problems with the Graham bill as currently drafted, at least some provisions of which cause a lot more mischief than Ben’s posts suggest. (Some of these concerns have already been raised by Daphne Eviatar here and here.) But lest readers think that my concerns with the Graham bill can be “fixed” simply by addressing the critiques outlined herein, I conclude with a more general explanation for why, in my view, detention legislation like the Graham bill is both unnecessary and unwise. In short, my specific objections to S. 3707 notwithstanding, better drafting won’t—and can’t—solve my real concerns.

Because of the unusual length of this post, I begin with a roadmap: Part I summarizes the current version of the Graham bill. Part II offers my specific objections to it. Part III lays out the more general critique of such “framework” detention legislation.

I.  The Graham Bill in a (Rather Thorough) Nutshell

As currently drafted, the Graham bill would create new 28 U.S.C. § 2256, titled “Habeas corpus review for certain unprivileged enemy belligerents.” Section 2256(a), the definitions provision, defines “unprivileged enemy belligerent” (those who can be detained under the statute) as anyone other than a privileged belligerent (curiously, the bill doesn't really deal with detaining any of them) who “(A) has engaged in hostilities against the United States or its coalition partners; (B) has purposefully and materially supported hostilities against the United States or its coalition partners; or (C) was a member of, part of, or operated in a clandestine, covert, or military capacity on behalf of the Taliban, al Qaeda, or associated forces.” This definition is basically a carbon copy of the controversial definition of who can be tried by a military commission under the Military Commissions Act of 2009 (see 10 U.S.C. § 948a(7)), but I’ll come back to that point later. 

New § 2256(b) “reaffirms that the United States is in an armed conflict with the Taliban, al Qaeda, and associated forces and that those entities continue to pose a threat to the United States and its citizens, both domestically and abroad.” (emphasis mine). As such, the bill authorizes detention of anyone involved in that conflict, “regardless of the place of capture, [and] until the termination of hostilities.” New § 2256(c) deals with “jurisdiction and venue,” although in addition to vesting exclusive jurisdiction in the D.C. district court (even if the detainee is moved somewhere else), § 2256(c)(2)(B) purports to bar the exercise of that jurisdiction over any claim other than a “challenge to the legality of the continued detention” of the petitioner, including claims “relating to the detention, transfer, treatment, trial, or conditions of confinement of the covered individual or any other action against the United States or its agents.” 

New § 2256(d) deals with the issue of “organizations considered associated forces,” and provides procedures through which the detention authority provided by § 2256(b) can be brought to bear against those affiliated with groups other than al Qaeda and the Taliban, procedures that even Ben finds somewhat troubling given that "[a]s written, the provision would do nothing to prevent a secret, sealed notice to the court and no notification of Congress at all, when the government effectively expands the legal war by interpreting the AUMF as covering a new group."

Section 2256(e), titled “Procedures,” is the most intricate provision in the bill. In particular, § 2256(e)(1) deals with burdens and evidentiary presumptions, including the language to which Daphne objected (e.g., in § 2256(e)(1)(E)) that creates strong presumptions both with regard to a detainee's training and his membership in the relevant organizations that can only be overcome with "clear and convincing evidence."

Section 2256(e)(2) governs discovery, and § 2256(e)(3) deals with witness production. There’s interesting stuff in both (and, as Ben notes, the bill is more favorable to the government than the current case law), but I’m going to basically skip over them here given the more important fish to fry elsewhere. Section 2256(e)(4) deals with the admissibility of statements alleged to have been obtained involuntary, including those obtained through torture or other coercion. Curiously, other than the bar on admitting statements obtained through torture, the provision otherwise allows admissibility so long as a statement was either voluntary or "was made incident to lawful conduct during military operations at the point of capture or during closely related active combat engagement, and the interests of justice would best be served by admission of the statement into evidence."

Section 2256(e)(5) provides for the “reliability” of statements against interest; § 2256(e)(6) governs attorneys representing detainees in habeas proceedings; and § 2256(e)(7) provides procedures for the use of video testimony. Again, although there is fodder in each of these, I’m going to leave it to the reader to delve into their specifics. 

Most commentary on the Graham bill thus far has completely overlooked new § 2256(f), which all-but precludes two additional classes of habeas claims: those seeking to collaterally attack military commission proceedings, and those seeking to block transfers to third-party countries. To that end, § 2256(f)(1) requires a stay of any habeas petition brought by a detainee against whom military commission charges have been sworn; § 2256(f)(2) creates significant exhaustion requirements for military commission defendants who have been convicted by a military commission before they can use habeas; and § 2256(f)(3) requires a stay of any habeas petition in a case where the executive branch is making "good faith efforts" to secure the transfer of a detainee who has been "designated" for transfer or release. 

Finally, § 2256(g) deals with second or successive applications for habeas relief, and § 2256(h) governs the release of detainees, including § 2256(h)(1), which bars courts from ordering release into the United States, § 2256(h)(2), which leaves up to the Secretary of Homeland Security to transfer a detainee who has been ordered released. I may be skipping other relevant provisions, but it's a dense 23-page bill, and I wanted to focus on the highlights.

II.  Problems With the Graham Bill

There are any number of problems with specific provisions of the Graham bill.  For ease of reference, I am going to break the issues discussed below into five major categories, in loose order of their structural-ness: (1) the identity of the enemy; (2) the scope of detention authority; (3) the bars on claims that might be protected by the Suspension Clause; (4) the evidentiary presumptions; and (5) the voluntariness of statements. I by no means suggest that the following is an exhaustive catalog; rather, it is just the issues that jumped out at me when reading through the bill.

A.  The Identity of the Enemy

One of the most remarkable things that the Graham bill does--in two separate places--is provide for the potentially fundamental expansion of the scope and nature of the armed conflict Congress authorized in the September 2001 Authorization for the Use of Military Force (AUMF). Just to remind ourselves, here's the critical authorizing language enacted nine years ago next week:

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 11, 2001, or harbored such organizations or persons, in order to prevent any future acts of international terrorism against the United States by such nations, organizations or persons.

Whatever else might be said about the AUMF, the requirement that the organizations have a direct link to the September 11 attacks is manifest, and has routinely been enforced by the D.C. courts in detention cases (as manifested perhaps most strongly in the D.C. Circuit's Parhat decision).

The Graham bill provides for the silent extension of the conflict in two distinct places: First, § 2256(b)(1), the "Statement of Authority," "reaffirms that the United States is in an armed conflict with the Taliban, al Qaeda, and associated forces," (emphasis added), regardless of whether those associated forces had anything to do with 9/11 whatsoever. Second, and relatedly, § 2256(d), which deals with "Notice of Organizations Considered Associated Forces," requires district courts to give "utmost deference" to the government's "notice" that "a particular organization" is "among the associated forces of the Taliban or al Qaeda." (A notice that could presumably be filed ex parte and under seal).  In other words, the government can simply assert (behind closed doors) that a group is associated with the Taliban or al Qaeda, and that assertion is not just effectively unreviewable, but is itself the basis for detaining individuals affiliated with that group (whether or not they have any ties to the Taliban or al Qaeda).

Indeed, the bill is dead silent on what standard the government should even use in ascertaining whether a particular organization is "associated with" the Taliban or al Qaeda. Such authority could hypothetically be used to work a dramatic extension both of the class of individuals who can be detained without charges, and, relatedly, the scope of the conflict Congress authorized nine years ago. Indeed, for all of the current debate over whether the AUMF authorizes the use of drones far from active military theaters, including in Somalia, Yemen, and other parts of the world, this provision could well be viewed as sanctioning such authority, so long as the targets were affiliated with "associated forces." 

B.  The Scope of Detention Authority

As noted above, the Graham bill provides something vital that the AUMF did not: a specific definition of who can be detained (new § 2256(a)(6)). So far, courts have been relying on the vague language of the AUMF quoted above, and the assumption derived from the Supreme Court's opinions in Hamdi and Hamdan that the scope of authority under the AUMF can be understood at least largely by reference to the laws of war. (This was the heart of the fight over the D.C. Circuit's Al-Bihani decision, as I explained here.).  The Graham bill provides specific detention criteria, borrowing rather expressly, as I noted above, from the 2009 MCA. Under the bill, the government can detain individuals who are not "privileged enemy belligerents"; and (1) have engaged in hostilities against the U.S. or its coalition partners; (2) have "purposefully and materially supported hostilities against the United States or its coalition partners," or (3) were members of, parts of, or "operated in a clandestine, covert, or military capacity" on behalf of the Talbian, al Qaeda, or associated forces. 

Assuming, for the sake of argument, that (1) and (2) are basically consistent with international law (I recognize that this may not be a satisfying assumption, especially re: (2)), the real problem is (3), because the laws of war are hardly clear that being a "member" or a "part" of a group like al Qaeda or the Talbian automatically makes one a "belligerent" subject to detention under international law. I don't mean to assume the debate over whether international law meaningfully constrains the scope of Congress's power to authorize detention without charges.  But for those, like me, who believe that it does, this is a substantial defect in the Graham bill (let alone in the 2009 MCA, which raises the separate and additional question of whether such individuals can be tried by a military commission).

Nor is it any consolation to argue, as Ben does, that this language merely codifies existing case law. There is dicta in the D.C. Circuit's original Al-Bihani decision to the effect that this is the scope of the government's detention authority, but (1) it's only dicta; and (2) it is not even a little clear how the 2009 MCA, which was not in any way addressed to detention, could have expanded the detention authority conferred by the AUMF. Thus, for better or worse, the Graham bill would dramatically enlarge the scope of who could be detained as compared to the status quo (or, at the very least, would codify the erroneous dicta in Al-Bihani), and in a manner that might well be inconsistent with international law. [In addition, wrangling over what it means to be "part of" al Qaeda is itself an issue that would surely create a similar range of decisions from the district court as those that we've seen under the AUMF. So the Graham bill replaces one ambiguous set of terminology with another.]

Finally, although it's a bit of a tangent, the Graham bill asserts that detention authority exists "until the termination of hostilities," and specifically disavows any need on the government's part to show that the detainee represents a continuing threat to the national security of the United States.  Although I, for one, do not think such provisions are inconsistent with international law (assuming that the individual is properly detained as a belligerent), there may be due process issues (to the extent the Due Process Clause applies), and there are certainly policy concerns, as Ben documents here. Periodic review should be a staple of any long-term detention regime, and yet the Graham bill's only response to this concern is to constrain the scope of second-or-successive applications for habeas relief in new § 2256(g). (For more on this problem, see Peter Margulies' thoughtful analysis.)

C.  Barring Claims that Might Be Protected by the Suspension Clause

Although the Graham bill has been marketed almost entirely as being geared toward the post-Boumediene "detention" cases, it includes several provisions that would constrain judicial review of other claims, and which may thereby implicate the Suspension Clause as applied to the Guantanamo detainees. Ben emphasizes in one of his posts that the Graham bill's "limited impact" is "one of its virtues." As evidence, he suggests that "[i]t does not in any way impact the scope of federal court jurisdiction to entertain habeas claims." (emphasis his). Quite to the contrary.

First, new § 2256(c)(2)(B) bars the use of habeas to litigate claims "relating to the detention, transfer, treatment, trial, or conditions of confinement" of the petitioner other than pure challenges to the legality of continued detention. It is hardly settled that these claims all fall outside the scope of the Suspension Clause, especially those relating to transfer. Even Kiyemba II, the problematic D.C. Circuit decision that rejected the ability of the D.C. courts to enjoin a detainee's transfer until the detainee had a hearing and an opportunity to be heard, turned on the government's assurance on the merits that it would not transfer detainees to countries in which they credibly feared mistreatment.  To take away jurisdiction over such a claim regardless of whether the government has made such assurances raises serious Suspension Clause concerns, since it bars access to habeas even where a detainee can prove that he credibly fears torture if returned to his home country (and that such a transfer would therefore violate both domestic and international law).

Second, new § 2256(f)(1) imposes a mandatory stay of any habeas petition brought by a detainee against whom military commission charges have been sworn, even though the Supreme Court has repeatedly recognized (most recently in Hamdan) that defendants before military tribunals have a right to contest their amenability to military trial before the trial takes place. As Justice Stevens put it in Hamdan, "we do not apply Councilman abstention when there is a substantial question whether a military tribunal has personal jurisdiction over the defendant." To be sure, nothing in Hamdan (or any other Supreme Court decison) suggested that such a right not to be tried is protected by the Suspension Clause, but as I've argued elsewhere, there is at least a colorable (if not substantial) line of reasoning supporting that conclusion. Indeed, there would be little point in protecting the ability of military defendants to collaterally invoke their right not to be tried if such a right had no constitutional foundation.

Third, proposed § 2256(f)(2)(D) constrains the substantive scope of post-conviction habeas petitions attacking military commissions to only those claims suggesting "that the commission was without jurisdiction." In other words, no habeas for non-jurisdictional constitutional defects at trial. At least since 1941, the Supreme Court has recognized a broader scope for federal habeas review of state-court convictions, and there are fairly strong arguments that at least some of that review may be constitutionally compelled. And as Justice Frankfurter repeatedly pointed out, whatever deference federalism might require the federal courts to show to convictions obtained in state courts, similar arguments are far less availing when the trial court was a federal military tribunal.

Fourth, new § 2256(f)(3) imposes a mandatory stay of any habeas case where the petitioner has been designated for transfer or release, and where the government is making "good faith efforts" to facilitate that result. In other words, as long as the government is, in good faith, trying to effectuate a detainee's release, the courts not just should, but must, stay their hand.

Especially with regard to these objections, I suspect one response will be that they are largely reflective of extant case law. Section 2256(f)(1) is basically the D.C. district court's decisions in the Hamdan and Khadr litigation; and § 2256(f)(3) reflects the practical result of the D.C. Circuit's decisions in Kiyemba I and Kiyemba II -- that there's nothing more for the courts to do once the government avers that it's doing everything it can to effectuate a detainee's transfer and/or release. To me, at least, such a rejoinder is unconvincing. For starters, there are serious problems with each of those decisions (especially Hamdan, Khadr, and Kiyemba II), and I'm hopeful (perhaps unjustifiably so) that the Supreme Court will eventually distance itself from the relevant conclusions contained within each of them.

But more fundamentally, it is one thing for courts to decide that discretion is the better part of valor, and that they should stay their hand while other events (be they military commissions or diplomatic negotiations) are underway. It is quite another for Congress to affirmatively bar the courts from acting, and in a manner that, as noted above, may go beyond existing precedent. In short, I don't think it is a sufficient defense of these provisions that they merely codify existing law.  Even if that were descriptively accurate (and I don't think it is), it assumes that existing law is correct--and final.

D.  Evidentiary Presumptions

I have somewhat less to say about the various evidentiary presumptions that the Graham bill creates, not because they aren't significant, but because Ben and Daphne have already discussed them at some length (Ben himself is critical of most of them, for reasons with which I basically agree). The shorthand, though, is that proposed § 2256(e)(1) creates strong presumptions supporting the government's authority to detain where there is a showing that the detainee received any "military-style" training from the Taliban, al Qaeda, or associated forces (back to those guys again), and an even stronger presumption against a detainee's withdrawal from his affiliation with that organization. (So strong that it cannot be rebutted after capture, even if, as in the Slahi case, the detainee turns state's evidence.) Section 2256(e)(1)(E)(2) requires a detainee to show withdrawal by clear and convincing evidence, something that will be extremely difficult for an individual detained at Guantanamo -- and without meaningful access to witnesses, or other potentially exculpatory evidence -- to do.

E.  Voluntariness

Finally, the Graham bill also redefines when statements can be admitted against a detainee in § 2256(e)(4). In particular, statements can be admitted so long as they were not obtained through torture or other cruel, inhuman, or degrading treatment; and (1) they were voluntarily given; or (2) they were made incident to lawful conduct during military operations at the point of capture or during closely related active combat engagement, and the interests of justice would best be served by admission of the statement into evidence." For the most part, I find this provision unobjectionable. My fear, though, is that this last provision could be broadly construed to admit against detainees any number of statements they make in the process of their capture, including statements that would not traditionally be admissible as excited utterances, statements against interest, and so on. Put simply, why isn't voluntariness enough, here? 

But the real trouble may come from new § 2256(e)(4)(D), which creates "a rebuttable presumption in favor of the voluntariness of statements against interest given before a Combatant Status Review Tribunal, Administrative Review Board, or comparable review board or as a result of treatment in compliance with the Army Field Manual." (emphasis added). In other words, any statement against interest not obtained through torture is presumed to be voluntary. Hmm...

F.  Concluding Thoughts on the Graham Bill

At the end of one of his posts defending the Graham bill, Ben poses the following question: "Do you object to legislative action that does nothing to alter the status quo but merely codifies the points the courts have already largely settled?" I've endeavored in the above paragraphs to explain the ways in which the Graham bill does in fact alter the status quo, or, at the very least, how the current status quo is itself problematic, whether as a matter of policy, legal constraints, or both. In that sense, then, Ben's question is misleading, because "merely codif[ying] the points the courts have already largely settled" isn't really (or only) what the Graham bill does. It also is worth emphasizing that, as with the "part of" language vis-a-vis the scope of detention authority, the very interpretive inconsistencies that supporters of the Graham bill invoke to justify such legislative reform will likely recur under the Graham bill, albeit in the context of different language. Some of the problems critics have identified just aren't problems; others are problems, but inherent ones. But, if we assume for the sake of argument that Congress could write a statute that did not raise the policy or legal problems charted above, would that be a good thing? That's the harder question to which I now turn.

III.  Structural Problems With Any Detention Legislation

In one sense, it is more than a little silly to object categorically to detention legislation. After all, we're only here because Congress passed a statute (the AUMF) that, as interpreted, authorizes the detention without charges of certain individuals. And so one might understandably assume that since detention authority is up to Congress, it makes perfect sense for Congress to fill in the rest of the puzzle, as well.

The reality, though, is more complicated. Habeas is at its core an intensively fact-bound inquiry into whether the jailer had the legal authority to incarcerate the prisoner. Yes, there are baseline rules that Congress can articulate (e.g., we're at war with Germany, not Australia), and evidentiary standards Congress can adopt (e.g., preponderance vs. clear-and-convincing), but it is difficult for a one-size-fits-all statute to otherwise deal with every contingency.  Should Congress really be specifying whether one night in a Taliban training camp is enough to make someone a "part of" that organization, as opposed to two? Does it make sense for legislation to specifically delineate the circumstances in which evidence should and should not be given particular weight? Or to precisely differentiate between the kinds of documents that should and should not be discoverable in individual cases? Are there reasons why Congress is better suited to make these decisions across the board than district judges, on the records of the individual cases in which they arise?

Justice Brennan once wrote that the Supreme Court has "consistently rejected interpretations of the habeas corpus statute that would suffocate the writ in stifling formalisms or hobble its effectiveness with the manacles of arcane and scholastic procedural requirements." At a fundamental level, this sentiment reflects the belief that habeas is essentially an equitable remedy, designed to be shaped by judges based upon the circumstances of individual cases, and not by legislators who can't possibly (and shouldn't be expected to) account for such hyper-specific variations.

But there's an even more pernicious problem with "framework" detention legislation, and it's one that's hardly new.  In his fascinating recent study of habeas practice in England from 1500-1800, Paul Halliday documented how Parliament's increasing "codification" of habeas actually weakened the writ, by cutting down the presumption that King's Bench had the sweeping common-law powers that it routinely exercised in habeas cases well into the 1700s. Even benign habeas legislation (a characterization that I do not believe the Graham bill deserves) had the effect in England of cabining the discretion of judges to adapt the remedy that habeas is designed to provide, a practice that I dare say American history has reflected.

This is not to say that there is no role for Congress here.  I freely concede, as I must, that the scope of detention authority is entirely Congress's baliwick, just so long as the legislature operates within constitutional limits. I believe that Congress should trust the courts to apply a general rule to specific facts, but I also recognize that, in the end, this question is, and must be, a threshold policy question for our legislators, that judges are then charged to implement. But the further afield Congress goes past identifying those who may be detained, the more it is wading into the traditional (and necessary) equitable discretion of habeas courts--the more it is assuming the job not just of lawmakers, but law-appliers and law-expounders. Either Congress must leave interpretive flexibility in the hands of the courts, or it must legislate with a scalpel. I have a hard time believing that the latter is truly the approach we should be sanctioning in an area so dominated by gray, even if that gray itself creates some uncertainty in the decisional law.

As those familiar with my work (and recent postings) should know, I am no fan of the current status quo in the Guantanamo litigation, especially the vast majority of decisions in detainee cases by the D.C. Circuit. As a result, it may seem strange that I am devoting so much energy to defending the power of the courts when I am so troubled by how those courts are (or, more precisely, aren't) using that power. But as problematic as some of the recent decisions have been, framework legislation like the Graham bill could do even more damage, on a far wider (and longer-range) scale. For better or worse, the Guantanamo cases are a short-term problem. But the notion that these kinds of decisions are for Congress, and not the courts, is a principle with the ability to transcend the current debbate, and to go to the very structural separation of powers between the branches.

As such, absent a far more convincing and compelling showing that the current, judicially-driven status quo is unworkable, and that the proposed legislation meaningfully fixes those shortcomings (indeed, many of the problems with current practice identified by Ben are either unaddressed, unresolved, or perhaps even exacerbated by the Graham bill), we should all be exceedingly wary of any legislation, from any quarter, purporting to have all, or even many, of the answers.