The State Secrets Privilege. I know little about this doctrine and defer to Bobby’s superior expertise. I will just make a simple point that will by now be familiar. The state secrets privilege, like the other rules we have discussed, reflects a tradeoff between liberty (or some other value at stake in a particular case) and security. The privilege allows the executive to maintain secrecy where publicity would aggravate a security threat, but by the same token enables the executive to engage in various forms of abuse without detection and serious review by the courts. It is tempting to say, and it is said often, that whatever other powers the executive should have during an emergency, at least its actions should be transparent, so that the public can evaluate them. If it must detain people without charges, for example, at least make it do so in the sunshine. Unfortunately, this response simply denies the premise that in some cases loss of secrecy can have devastating consequences, a premise that no one denies but everyone tries to minimize in arbitrary ways (typically by admitting only that troop movements might appropriately be secret, as though troop movements were a phenomenological category carved into the structure of the universe). The risk of executive abuse is the necessary cost, just as it is when one sets up a government in the first place. In this respect, I believe that our view is no different from Bobby’s. What I would add is only that while the privilege must surely be necessary during normal times, it ought to be broader during emergencies, for all the reasons that we have been discussing: when the threat to security increases, a higher risk of executive abuse needs to be tolerated. Of course, if the threshold is already extremely low during normal times, it might not be necessary to expand it during emergencies (but then the privilege should probably be more limited during normal times); or it might be that the degree of emergency is implicitly built into the threshold.
Alternatives. Julian’s post correctly notes that most legal academics think that the judiciary should play a significant role to restrain the executive during the emergency. But there is a third view, generally associated with Cass Sunstein, and Rick Pildes and Sam Issacharoff. On this view, courts should be deferential as long as Congress has authorized the executive action in question. The kick here is that congressional authorization that would not immunize executive action from judicial challenge during normal times does so in an emergency. I do not agree with their arguments, but I can see how many people would find them appealing. If we could reproduce something like the war cabinet system that has served Britain well, that would indeed be worth considering. But the U.S. system is nothing like the British system, and I do not think Congress or members of Congress or particular committees could operate as a kind of war cabinet. Still, the Sunstein/Pildes/Issacharoff proposals have a very significant merit: they make clear that the problem of emergency powers is just an institutional problem. It involves thinking about decision costs, error costs, and the relative capacities of the different branches of government. We all might have different ideas about how to weigh these factors, and once we discuss these different ideas, it becomes clear that the debate is empirical and (at a very low level) analytic, not ideological. Worries about tyranny, fascism, deontological principles, and the like, could be put in the same dustbin as the Federalist Papers.
Censorship. I do think that censorship raises the same issues as the other types of executive actions that we discuss, and that the tradeoff and deference theses apply with equal force. It is interesting to speculate about what would happen if the U.S. had the homegrown terrorism problem that currently afflicts Britain. Would U.S. authorities try as hard as the UK government to censor firebrand imams? It’s hard to say. I expect they would if the problem was serious enough, and I expect courts would defer. But, given U.S. traditions, the problem would probably have to be more serious than it currently is in Britain.
A general thought. One other thought, which has been provoked indirectly by a few of the posts and some of the comments, but mainly by other reactions I have seen from time to time: many people seem to think that the question of the proper tradeoff between security and liberty during emergencies, and the proper role of different government institutions, can be resolved by parsing doctrine. This is most definitely not our view. Given certain broad institutional constraints, the problem is mainly normative and empirical. The contributors to this debate who pour over opinions, statutes, and the constitutional text are often very impressive lawyers, but they are going about the problem the wrong way, and, inevitably, they end up supplying answers by smuggling in normative and empirical assumptions that they do not acknowledge and defend.
Kevin’s Questions. On the first, the book not being a brief for the Bush administration, the only question is whether the hypothetical second 9/11 attack sheds new light on the relative institutional capacity of the presidency. It might not; or it might show that the presidency is not strong enough; or it might show that the presidency is too strong. Everything would depend on the reasons for the attack’s success, and whether one thinks a more constrained president would have either prevented the attack or done less harm to civil liberties in the course of failing to prevent it. Suppose that during World War II, the German offensive in the Ardennes had succeeded and allied troops were driven off continental Europe. Would it have been proper for Congress and the courts to assert control over the executive on the grounds that FDR was incompetent? It would probably take something even worse than that to precipitate such a constitutional crisis.
On the second question, I can’t think of any off the top of my head, but that is probably just due to my ignorance. But one can think of hypotheticals. An extreme example I suppose would be a full-scale invasion of Iraq, designed to weed out some suspected al Qaede elements and to convert Iraq into a model Muslim democracy, a beacon for the Muslim-Arab world, one that would help eliminate the roots of the extremist ideology that resulted in the first 9/11. Although judges could do nothing about this hypothetical Iraq invasion, it would probably not deserve deference from Congress.
My thanks again to Opinio Juris, and Roger in particular, for organizing this symposium. I am sorry that I have not been able to respond to the many interesting comments.
I appreciate Adrian’s thoughtful response to my post on military detention, and would now like to shift gears to a distinct topic of at least equal current significance: the state secrets privilege (“SSP”).
We could have a whole symposium on this issue alone, no doubt. In fact, we had one a few months ago over at my usual blogging abode, National Security Advisors; I strongly recommend that exchange for anyone who wants to get down into the details of the SSP debate. I do not propose to rehash the entirety of that debate here, though, but instead simply to flag this topic as one in which the deference thesis is currently in issue in live litigation; anyone who has listened to last week’s oral argument in Hepting v. AT&T Corp, before a Ninth Circuit panel, will appreciate that there continues to be great uncertainty even among judges with respect to (a) the doctrinal details through which the SSP is implemented and (b) the separation of powers principles that should give rise to those details (see in particular the exasperated exchange between Judge Pregerson and Dep. SG Garre with respect to the deference judges owe in this context).
According to the Supreme Court’s 1953 decision in Reynolds, the substantive test for application of the privilege is whether disclosure of the information in issue poses a “reasonable” risk of harm to national security. As a formal matter, the existence of such a risk must be expressly asserted by the head of the agency or department with responsibility for that information (the DNI, for example). At the same time, the majority in Reynolds expressly rejected the proposition that the mere assertion of the privilege by the executive should bind the court; whatever deference the executive official should receive, it ultimately is the responsibility of the court to decide whether the reasonable-risk predicate truly has been satisfied.
Assuming that description of the doctrinal status quo is correct, would the deference thesis require a shift in the doctrine such that the executive invocation of the privilege becomes dispositive? I’m not sure that this is so—perhaps the thesis would cash-out in this context precisely where the doctrine already is—but if it is, then it seems to me that this would not be desirable. On one hand, the existence of at least some prospect for judicial review has a salutary effect: the possibility that a judge might reject an assertion of the privilege serves an important role in disciplining executive officials in their invocation of it, counterbalancing any temptation there might be to invoke the privilege in entirely unwarranted circumstances (something that I think would be rare, but which I would not rule out). Against that benefit, there is a potential cost: a judge might improperly reject a valid invocation of the privilege (something that I think would be rare, but which I would not rule out).
Does one consideration outweigh the other? Before answering that question, I think it is important to recall the relatively-forgiving substantive standard to which the judge holds the executive official: demonstration of a reasonable risk to national security should disclosure occur. Because that threshold is quite low, it is relatively difficult for a judge to plausibly assert that the executive official has not satisfied it (indeed, this is the cause of considerable criticism of current SSP doctrine). As a consequence, the potential for unwarranted rejections of the privilege is constrained relative to what might occur under a tougher predicate requirement. I can’t quantify the resulting costs of permitting second-guessing by the judge, and therefore can’t entirely follow through on this cost-benefit analysis. Still, this constraint on the judicial error variable gives me some reason to believe that the benefits are indeed worthwhile.
(Knowing that many readers are extremely critical of the SSP, by the way, I feel obliged to add at this point that I do think that there is room for creative measures designed to alter the SSP status quo in limited ways. I’ve argued elsewhere that Congress should consider creating FISC-like fora for litigation of certain claims that otherwise would be dismissed under the SSP, and that there might even be ways to provide expert assistance to judges engaged in the process of considering whether the reasonable-risk standard has been met in a given case.)
That probably will be my final contribution to this terrific discussion. I’d like to thank Roger and the rest of the OJ crew for having me, the other participants and commentators for their thought-provoking observations, and Eric and Adrian for making such a significant contribution to the literature.
Bobby Chesney’s post about military detention asks all the right questions and I agree with the first several steps in his analysis of those questions. I want to record some questions or quibbles about the later steps in his analysis, however. It is true that Hamdi baldly forecloses “indefinite detention for the purpose of interrogation,” but it is possible that the emphasis here should be on “indefinite,” rather than on “interrogation”, especially in the context of the Court’s discussion, in which the central worry seems to have been that detention could go on forever (whatever the grounds of detention). In any event, our point of course is to criticize Hamdi’s procedural framework for detention, although of course we approve of its holding that detention is substantively authorized by the AUMF. We don’t take Hamdi‘s procedural holding as a fixed point; its correctness is what we mean to question.
If Hamdi’s holding (dictum?) barring detention to acquire intelligence is correct, it would have to be because, as Chesney explains lucidly, there is some predictable skew in the government’s decisionmaking, such that it will detain too many people, inflicting too many costs for too little benefit. The benchmark for determining whether such a skew exists is not just the number of false positives and false negatives, but a weighted comparison of their costs and benefits. A false positive – here, detaining someone who has no useful intelligence – might be less costly overall than a false negative – failing to detain someone who does have useful intelligence. The former is of course more costly to the detainee, the latter to society; a rational and well-motivated government would weigh these costs impartially. The crucial questions, then, are the rates of false positives and negatives, the costs of false positives and negatives (which are not necessarily equivalent), the government’s incentives to conduct the calculus accurately, and whether the judges can improve upon the government’s decisionmaking.
I have not yet heard any account suggesting that if the government could detain in order to gather intelligence it would do so to excess, or that the judges could improve upon its decisionmaking, although I am open to thinking that there is such an account. I will say only that false positives are hardly costless to the government in this setting. One of the comments to Chesney’s post (by Geoffrey Corn) mentions several such costs, such as “distracting resources from more precisely focused intelligence gathering, and overwhelming the logistical capability to hold and care for such individuals.” Even if the government only cares about its own costs, not total social costs, it will have an incentive to keep down the number of false positives. Whether the incentive is adequate is difficult to say; by the same token, however, I see no basis for judges to be confident that it is not. Corn mentions that front-line military personnel may have an incentive to over-detain, but by doing so they are imposing costs on their superiors, who will be aware of the perverse incentive and will try to do something about it. If the problem is agency slack within the military – the commanders cannot fully control the line officers — the notion that judges can fix the problem through due process review strikes me as far-fetched.
As this will have to be my final post, I want to thank our commentators and my co-author for their interesting remarks, and thank Roger Alford for hosting this event.
Detention Policy, the Executive Interest in Intelligence-Gathering, and the Calibration of Procedural Safeguards
I’d like to steer the discussion toward the question of military detention for a moment.
Military detention has been and continues to be the subject of extensive litigation, and it therefore presents a series of occasions implicating the deference thesis. Eric and Adrian discuss the matter from several angles, including one that strikes me as particularly important: procedural safeguards (i.e., the mix of rights, procedures, and rules governing the process of determining whether a particular individual satisfies the substantive criteria for military detention).
In any fact-finding system, the applicable procedural safeguards reflect (even if only implicitly) an underlying judgment regarding the appropriate rates of false positives and negatives to be tolerated by that system, as well as a judgment regarding the mix of safeguards likely to produce such rates (all of which, of course, is but another way of stating the tradeoff thesis). From the perspective of the deference thesis, the question that arises is whether courts (or Congress) should second-guess executive determinations of the calibration most appropriate for military detention in the current security environment. Eric and Adrian note that the Supreme Court did precisely that in Hamdi, and they criticize the court for doing so (pp. 256-57). Because the question of procedural safeguards continues to be a point of special significance and sharp dispute, it seems to me that their argument on this point deserves close consideration.
As an initial matter, Eric and Adrian make clear that they concur in the need for a cost-benefit analysis regarding procedural safeguards in this context. They would leave that determination in the hands of the executive, however, in the absence of an explanation as to why judges will conduct a more accurate analysis. “One needs a theory,” they point out, “to explain why the government will weigh the relevant factors with systematic bias, as opposed to random error . . . .” (p. 256).
Hamdi does not offer such a theory (at least not expressly), and this prompts Eric and Adrian to identify and then critique an account that might explain the court’s approach: “One such theory . . . is that the government will just want to detain in perpetuity or convict before a military commission everyone the president alleges is an enemy combatant, so the procedural cost-benefit analysis will be distorted.” (Id.). They describe this account as “mystifying,” however, explaining that “[t]here is no reason to think that the executive would benefit from an excessive detention or conviction rate, or that political constraints would permit the executive to implement such a preference in any event.” (Id.) (emphasis added). Later on the same page they explain that the government has a strong interest in accurately sorting “real enemy combatants from those who are swept up in the fighting by mistake,” and that this interest involves the desire to “incapacitate and deter terrorists and attackers . . . .” (Id.)
If it were the case that the executive interest in detention were limited to the incapacitation of detainees themselves and the deterrence of others, I would find this to be a fairly persuasive line of argument. But it seems to me that the executive interest in detention goes beyond incapacitation and deterrence, encompassing also the powerful interest in gathering intelligence about al Qaeda and the like. In light of that interest, there is at least some reason to believe that the range of persons whom the government would like to detain might exceed the range of persons properly subject to detention pursuant to the traditional armed conflict rule relating to internment for the duration of hostilities. Put another way, the government’s interest in gathering intelligence may lead it to wish to detain (and then interrogate) persons whose detentions might be difficult to justify under conventional understandings of the scope of internment power.
Let’s assume this is so. So what? We need two additional premises before we can adopt this as a viable theory for why the executive might indeed preference an excessive rate of false positives in the detention process.
First, we need to assume that the government will not or cannot invoke intelligence-gathering, standing alone, as a detention predicate. That seems a safe assumption, actually, in light of the express language in Hamdi foreclosing that option (noted by Eric and Adrian at p. 254).
Second, we need an account that explains why heightened toleration for false positives might facilitate the intelligence-gathering interest. Is there such an account? Possibly so. Much of the information that would be most helpful in unwinding al Qaeda and comparable terrorist networks lies in the hands of individuals who might not easily be described as actual members of such groups, let alone individuals who engaged directly in hostile acts. Financiers, document forgers, and other components of the logistical back office for terror networks serve a critical function, and may be positioned to provide tremendously useful information about other parts of the network. Some such individuals can of course be categorized as sufficiently associated with a hostile force so as to justify detention even on a relatively strict reading of traditional detention criteria. Others, however, present a more difficult case. The weaker the procedural safeguards involved in the process of determining detention eligibility, however, the easier it becomes to finesse such distinctions (and the more intel-rich detainees in theory might come into long-term custody).
Certainly this account provides an explanation for the importation of “material support” concepts from domestic criminal law into the military detention context. Whether it also provides a plausible account for questioning executive decisions relating to procedural safeguards is less certain, but worth considering.
In Terror in the Balance, we put aside the view that there is an absolute moral prohibition on coercive interrogation necessary to save third-party lives. For one thing, we claimed, it is very hard to find moral philosophers who defend that view; most waffle, in the end, by adopting some variant of the view that there is a “catastrophe” exception to the moral prohibition, so that if enough lives are at stake a utilitarian override kicks in. Roger Alford, our gracious host, suggests (among other things) that we radically understate the appeal of absolutism. He points out that the McCain Amendment and various sources of international law create an absolute legal ban on “torture,” of which coercive interrogation is a subset.
But this point conflates the moral question with the legal question. There is no doubt that most people believe that coercive interrogation should be illegal, and indeed it is; but they also believe that it would be morally permissible or even obligatory for officials to torture, in some vaguely defined set of extreme circumstances, at least if those officials openly take responsibility for their actions and throw themselves on the mercy of juries and the public. McCain himself, in an article in Newsweek, wrote that torture should be illegal, but that in an “urgent” ticking time-bomb case “an interrogator might well try extreme measures to extract information that could save lives. Should he do so, and thereby save an American city or prevent another 9/11, authorities and the public would surely take this into account when judging his actions and recognize the extremely dire situation which he confronted.” In this, our guess is that McCain speaks for many Americans. (In the Pew Research Center poll that Roger links, “15 percent of respondents said that the use of torture against suspected terrorists in order to gain important information can often be justified, while 31 percent said it is sometimes justified, 17 percent said it is rarely justified and 32 percent said it’s never justified.” So despite the legal ban, 63 percent believe that torture is at least sometimes justifiable; we read this to mean morally justifiable.)
We label the view that McCain defends the OAF view — “Outlaw and Forgive” — and our objection to it is not that nobody holds it, but that it is bad in various ways: self-defeating, unstable, and socially undesirable. There is no need to repeat those points here. Conceptually, however, the undoubted legal prohibition on coercive interrogation does not show that anyone holds the absolutist moral view. Law and morality are not coterminous, in general or in the debate over coercive interrogation.
I agree that the concerns that Bobby Chesney identifies are real and important. There are no answers at the level of theory; the scope and level of deference must be worked out at the level of practical politics. In practice, as we have seen, the president (and presidents generally) press for maximal powers where they think they need them, subject to political constraints. President Bush has not argued that his commander-in-chief power gives him the right to dictate educational policy because such an argument is a loser politically as well as legally. The courts defer with respect to some actions and not others. Presidents often acquiesce when courts refuse to defer, but sometimes they put up varying levels of resistance—appealing up the chain, or jurisdiction-shopping until they get a better result, or exploiting loopholes, or buying for time, or in rare instances (FDR, Lincoln) disobeying or threatening to disobey judicial orders. Public and elite responses to the performance of the relevant actors gradually determines the practical limits on presidential, judicial, and congressional action. I’m afraid we don’t have anything illuminating to say about how boundaries should be determined in practice, or how context-specific deference ought to be.
But we do want to avoid the legalistic impulse to try to determine in advance what the rules should be. (Again, this is the impulse behind Ackerman’s Emergency Constitution.) There are obvious benefits from having rules stated in advance, but the rules/standards literature makes clear that there are costs as well. Emergencies are not like the revenue-generating behaviors that are regulated by the tax code. Because it is hard to anticipate the next emergency, rules determined today will inevitably be poorly suited to the emergency that occurs tomorrow. On balance, the unpredictability of emergencies argue in favor of general standards of conduct rather than rules.
On a somewhat related issue, some of the comments might give readers a misleading impression that we take a dichotomous view: that the choice is between presidential dictatorship or not, and we opt for the former. Bobby Chesney correctly notes that the real debate is about the proper location on a continuum. To pick silly numbers for clarity, suppose 0 is pure executive government, 100 is pure legislative government, and 50 is some mix of deference and congressional/judicial oversight. We do not argue for 0, nor does anyone argue for 100. To pick more silly numbers, suppose that on most issues in normal (non-emergency) times the system is at 50. During emergencies, it typically goes down to (say) 20. Some civil libertarians seem to argue for, say, 40 or even 60 during emergencies. We want to say that 20 seems right—or, more precisely, there is no reason for thinking that 20 is wrong. Civil libertarians make a series of arguments that 20 is too low—the panic argument, etc., as we noted earlier. In trying to refute these arguments, we are not committing ourselves to 0; we are committing ourselves to 20 for the duration of the emergency.
A complicating factor, which I referred to in an earlier post, is the existence of other trends—technological, cultural, geopolitical—that affect the optimal location on the continuum. So it may be that as weapons become cheaper, smaller, and more destructive, we will have to reconcile ourselves to a long-term decline from 50 to, say, 40. On the other hand, if foreign political extremism fades (as it has in the past, and will surely do again), the optimal point could rise from 50 to some higher number.