Conventional wisdom holds that our ongoing war against jihadi groups is utterly unlike a war against a traditional state adversary. In my previous post, I argued that this is at least partially untrue: Jihadi groups combine “top-down” elements that resemble, to some degree, hierarchically organized state adversaries, and “bottom-up” elements that more closely resemble a decentralized mass movement.
Yet our current model of how this war ends under law fails utterly to account for this. Recall that the AUMF “does not purport to limit the time period in which the President can act.” As I noted in my first post, longstanding precedent holds that that in such situations, the courts will look to a “public act of the political branches,” such as a treaty, congressional Joint Resolution, or presidential proclamation, to fix the conflict’s end date. The orthodox view of conflict termination in “AUMF war” is thus that it will end on some future date when an act of the political branches declares the enemy to be defeated. Until that presumably distant day, war powers will continue on in full force.
There are several problems with this binary, “on/off” view of the end of the war. The problems arise, not because it is descriptively incorrect as a matter of existing law, but because it is a poor construct around which to organize a conflict of this nature.
The most frequently discussed problem is indefiniteness: If there is to be one legal endpoint to this war, it will almost certainly be very far in the future. And the indefinite exercise of constitutional war powers fits uncomfortably within our constitutional order.
One solution to the indefiniteness problem is a statutory sunset: Revising the AUMF so that, absent congressional reauthorization, it would expire on a specified date. Sunset provisions are popular among the cognoscenti: The draft ISIS AUMF proposed by the Lawfare brain trust in 2014, the Obama Administration’s February 2015 proposal, and the AUMF proposed last year by Rep. Adam Schiff each included a three-year sunset.
Sunsets address the indefiniteness problem. They also have the virtue of requiring our elected representatives to periodically reassess whether the war should continue, and if so, what its appropriate scope should be. Such reconsideration is valuable not merely as an abstract matter of healthy democratic governance, but because courts often treat longstanding executive practice unchallenged by Congress “as a gloss on ‘executive Power’ vested in the President.” Put simply, by sleeping on its power and responsibility to update the terms upon which this war is waged and instead allowing the Executive to determine the war’s scope, Congress risks gradually surrendering that power over time.
On the other hand, sunsets place the federal government’s most elemental function—protecting the public from external enemies—at the mercy of the same “cliff”-driven politics that have produced so many legislative debacles in recent years. I think that this is a bad and dangerous idea; others may deem indefinite war powers the greater concern, and thus may be willing to risk seeing the President’s statutory authority to fight terrorists shudder to an abrupt halt.
A subtler, but equally problematic flaw of an AUMF sunset is the risk that it might backfire, inadvertently further exposing Congress’s weak control over executive warmaking. What would happen if the AUMF were allowed to sunset with American troops still in the field? A President might (indeed, I suspect, would likely) continue to prosecute the conflict pursuant to his or her inherent Article II authority to defend the United States from external attacks.
The final problem with a sunsetting AUMF is especially important for our purposes: It wouldn’t address the fact, discussed in Part II of this series, that not every aspect of the conflict against jihadi groups will terminate on one single date in the future. A sunset, like the current model of how the “AUMF war” ends, is binary: the war is either on, or entirely off.
Yet a single legal end date—whether a date certain in three years or a date uncertain in the distant future—is an ill fit for this conflict, for several reasons. Perhaps the most obvious is that the conflict inaugurated by the AUMF is now several conflicts: a geographically-bounded counterinsurgency against the Taliban in Afghanistan; a man-hunting campaign against the remnants of “core” al Qaeda in the Af-Pak borderlands; a pitched battle with ISIS in Iraq and Syria; and a global campaign against al-Qaeda and now ISIS affiliates in Yemen, Somalia, and elsewhere.
A second reason, which I discussed in Part II of this series, is that jihadi groups are a combination of “top-down” elements that bear some resemblance to traditional state adversaries, and “bottom-up” elements that look more like a leaderless movement. Many of these groups, like states, have hierarchical organizational structures and are based in identifiable geographic sanctuaries. Yet, unlike state adversaries, these groups’ foot soldiers can’t be relied upon to stand down when their leaders are defeated or decide to moderate their groups’ aims. There is no reason to believe—indeed, there is overwhelming reason to doubt—that the centralized and decentralized elements of the conflict will end at the same time.
The way out of this conceptual conundrum is that the various legal authorities encompassed by the AUMF can be disaggregated and given authority-specific termination mechanisms that match their features.
What might this approach look like?
First, it’s important to separate those elements of the conflict whose termination should be determined by the executive or legislative branches, and those elements whose termination requires judicial fact-finding. Put differently, a conceptually sound approach to conflict termination in the “war on terrorism” requires us to isolate and address separately the top-down and bottom-up elements of the conflict.
Whether an individual jihadi remains a threat to the United States depends less on the intentions of his leader or organization than on his or her own present state of mind. Thus, it logically requires individual fact-finding rather than an overarching political determination. (We’ll return to this later.)
By contrast, whether a given jihadi group has been defeated, or whether hostilities have ended in a given geographic area, are quintessentially political determinations, best left to those responsible for the national defense. As the Supreme Court explained in Baker v. Carr:
[L]ack of judicially discoverable standards and the drive for evenhanded application may impel reference to the political departments' determination of dates of hostilities' beginning and ending.
As an aside, this one reason why at least two D.D.C. judges’ apparent belief that a habeas court should assemble a “record” on whether hostilities continue in Afghanistan is so unfortunate. Whether hostilities continue is a “legislative fact,” and judicial finding of legislative facts tends to produce “an uneven mixture of a priori conjectures and partially informed guesses, with occasional factual investigations of varying depth.” Judicial fact-finding mechanisms are designed to discover particular facts that pertain to the parties, not for making systemic judgments that call for policy expertise and discretion. The black-letter rule that courts needing to fix the end date of hostilities look to a “public act of the political branches” exists for a reason.
But acknowledging the need to defer to a public act of the political branches doesn’t tell us what that public act should be. The current AUMF leaves it entirely undefined—but Congress could choose to do otherwise.
Wartime statutes passed during World War I included a wide array of different termination provisions. Some provided that they would terminate upon the signing of a peace treaty; others, upon a presidential proclamation that the war had ended; some contained date-certain sunset provisions; still others provided for termination upon a specific presidential finding that the authority was no longer needed.
Similarly, World War II-era Congresses passed dozens of statutes intended to have only wartime effect, with an equally dizzying array of termination provisions. Some used vague formulations like “in time of war” or “for the duration of the war,” which, as Attorney General Tom Clark noted, effectively entrusted their end to the President’s unreviewable discretion. But many others had more specific termination triggers: some featured date-certain sunsets; some had sunsets with an option for earlier termination by the President or Congress (note the potential Chadha problem here); some were tied to the end of particular national emergencies declared by the President earlier in the war.
To summarize: Congress need not treat the AUMF war as one phenomenon with one date. Rather, it can and (if it ever updates the 2001 AUMF) should break it up into discrete components, and assign an appropriate “decider” to determine when each one ends.
For example, the authority to use military force to fight the Taliban in Afghanistan, to use air power and special forces to fight ISIS in Iraq and Syria, and to conduct DOD drone strikes against al Shabaab in Somalia should, logically, end at different times, when the discrete phenomena at which they are directed end. Several Lawfare writers have proposed such a “retail” approach to the geographic and organizational scope of the AUMF for purposes of commencing hostilities in new areas.
The present AUMF addresses these varied problems by crudely conferring undifferentiated and indefinite global war powers. Congress could refine that broad grant into a basket of of narrower authorizations equipped with customized termination provisions. These could include (i) authority-specific sunsets, where Congress is especially concerned about mission creep (for example, in Syria), or (ii) clauses providing that a given authority will terminate upon a presidential finding that a specified contingency has occurred, that a particular threat no longer exists, or that the statute’s operation is no longer necessary.
Congress has already used authority-specific sunsets in a number of post-9/11 counterterrorism statutes, including the USA PATRIOT Act and the FISA Amendments Act. The second possibility, a clause requiring the President to “switch off” an authority when he or she finds a certain condition met, is less intuitive but is also well established. Many World War I and World War II-era statutes used such termination triggers, and they are relatively common in national-security-related statutes. For example, the statute in the iconic Curtiss-Wright case gave the President discretion to activate an arms embargo upon a finding that it would “contribute to the reestablishment of peace.” The International Emergency Economic Powers Act similarly permits the President to toggle on or off various financial sanctions upon making specified findings.
“This authority shall terminate upon a finding by the President that ISIS has been defeated in Iraq and Syria” is, admittedly, a softer constraint than a date-certain sunset. (I think this one of its virtues; others less concerned about obstructing the President’s conduct of hostilities might disagree.) Yet that does not mean that it is no constraint. The existence of a statutory criterion would compel the President to publicly defend his decision not to invoke the termination provision if the specified contingency had, even debatably, come to pass. This provides some—but admittedly not all—of the democratic-accountability benefits of sunsets in a manner that is far less disruptive.
Whether an authority-specific sunset or a presidential “toggle” is appropriate for terminating a given authority would logically depend on how concerned Congress would be about allowing it to continue indefinitely, balanced against the importance of executive freedom of action.
Overall, this approach would allow Congress to better control and tailor the war powers it confers in the “war on terror,” while avoiding the inaccuracy of tying a vast array of legal authorities to a single, overbroad termination date—whether fixed by a hypothetical “act of the political departments” at some point in the distant future or by a blanket sunset provision.
That leaves us with the “bottom-up” elements of the AUMF war—in particular, detention of captured jihadis. We know from Hamdi that the AUMF authorizes law-of-war detention “for the duration of the relevant conflict.” Some district-court judges have taken this to mean that detention must cease upon the end of hostilities in Afghanistan. As I argued in my previous post, this isn’t what Hamdi held, nor is it a sensible yardstick for determining whether the “relevant conflict” has ended for an international jihadi for whom Afghanistan is merely one front among many in a wider war.
So if the end of U.S. combat operations in Afghanistan is the wrong metric, then how would one identify the end of the “relevant conflict” for determining when to release an international jihadi? The uncomfortable truth is that there is no single event—certainly not the end of U.S. combat operations in Afghanistan—that would be a meaningful end-signal for these purposes. Marc Sageman has intriguingly compared the structure of “global neo-jihadi terrorism” to “the structure of anarchist terrorism that prevailed over a century ago, when no such global coordinating committee was ever found despite contemporaneous belief in its existence.” There may be local commanders who command temporary allegiance, but the fighters are ultimately committed to the cause. Releasing hardened fanatics at the end of hostilities in Afghanistan while the broader conflict with al Qaeda and its ideological kin rages elsewhere would be folly.
There’s an interesting corollary here, which Ben Wittes discussed in Detention and Denial. Using the traditional law-of-war “end of the conflict” yardstick for releasing detainees is not inapt solely because it is potentially too lenient; it is also potentially too stringent. The standard theoretical model of POW detention automatically permits detention until the end of the relevant conflict because the soldier is presumed “bound by an allegiance which commits him to lose no opportunity to forward the cause” of his sovereign. That principal-agent assumption does not, we have seen, necessarily hold in the case of the international jihadi, who is attached to his commander or organization by ideological affinity rather than compelled to fight by the coercive power of his home state.
Put simply, each jihadi is fighting his own war. And this means that when the “relevant conflict” ends for detention purposes is logically an individualized determination—or, put differently, an “adjudicative fact” specific to “a particular case.” And such adjudicative facts—unlike legislative facts, such as whether the Taliban continues to pose a threat—are indisputably the stuff of judicial fact-finding:
The usual method of establishing adjudicative facts in through the introduction of evidence, ordinarily consisting of the testimony of witnesses.
How, then, should our law account for the need for individualized adjudication for whether a given law-of-war detainee remains hostile to the United States?
The answer, at least under current law, is not for the courts to undertake it themselves. The D.C. Circuit has already held that the government’s “authority to detain an enemy combatant is not dependent on whether an individual would pose a threat to the United States or its allies if released but rather upon the continuation of hostilities.”
Yet in Boumediene, Justice Kennedy hinted that the potentially indefinite duration of the AUMF war might force the Court to step in and define “the outer boundaries of war powers.” The Court is clearly uncomfortable with indefinite executive detention at Guantanamo Bay, and that discomfort has spurred two D.D.C. judges to adopt a less-deferential posture toward executive assurances that the conflict continues.
The concern should be is that this discomfort will ultimately impel the Supreme Court to go where courts should not—second-guessing the political branches’ better-informed determination that a threat persists and that military force remains warranted. Alternatively, one could imagine a fed-up Supreme Court imposing a novel, judicially fashioned continued-dangerousness requirement on law-of-war detention of captured terrorists.
It would be preferable for Congress to fashion such a regime by statute than for it to hand such sensitive policy determinations over to the courts by default. Ben and others have written in detail about what such a statute might contain, so I won’t belabor the possibilities here. Such a statute would not be an exceptional outlier in the American legal tradition; indeed, for those concerned, it might, in certain cases, produce outcomes more lenient than what law-of-war detention or even criminal trials for offenses like conspiracy and material support might produce.
Interestingly, the Periodic Review process at Guantanamo has, in one respect, accomplished as a matter of executive discretion what the current AUMF does not: It accounts for the fact that that the conflict can end for an individual detainee before it ends for the organization that sent him. Unfortunately, it does not—and presumably could not—lawfully account for the fact that the conflict might end for a fanatical jihadi after it ends for the organization that sent him. Note that whether or not this conceptual alignment is desirable does not depend on one’s view as to whether a given detainee remains too dangerous to release.
One important implication of the Periodic Review process is that its existence should, in theory, blunt Boumediene’s threat to step in and define “the outer boundaries of war powers” in response to the prospect of indefinite law-of-war detention. A preventive counterterrorism-detention statute should pass constitutional muster if it includes “adequate means … of insuring both the accuracy of individual detention judgments and the necessity of those detentions.” And both of those features already exist: The existing habeas process evaluates the former, while the periodic review process evaluates the latter. (Some might argue, however, that an internal Executive Branch process is not an adequate means of ensuring the necessity of detention.)
This should influence judicial analysis of the lawfulness of long-term preventive detention at Guantanamo because the President’s “discretionary executive authority” over captured enemy combatants can be “properly exercised … to avoid serious constitutional doubt.” The executive branch’s discretionary imposition of a necessity requirement upon law-of-war detention could thus insulate the existing detention regime from any conceivable constitutional challenge, even without a subsequent statute establishing such a necessity requirement.
Of course, if the first transfers of Yemenis to Saudi Arabia portends a significant reduction in the “un-triable, un-transferable rump population at Guantanamo,” many of whom are Yemenis, the law governing these detentions becomes largely an academic question.
Some might respond that all of this is an academic question; why worry about the end of a conflict whose end is nowhere in sight? There are several reasons: To act before indefinite detention provokes courts to step in and impose judicially improvised, ill-conceived restrictions on the Executive Branch’s conduct of the war. To address legitimate concerns about indefinite war powers. To reassert Congress’s control over the scope of an armed conflict it authorized long ago, under different circumstances, with a much smaller set of enemies and geographies in mind. And because while this may be an academic question now, it won’t be forever.