Editor’s Note: This is the second entry in a symposium Just Security is hosting in conjunction with this week’s release of Power Wars: Inside Obama’s Post-9/11 Presidency by Charlie Savage. The series includes posts by Oona Hathaway and more to come.

In an article in the New York Times last week, Charlie Savage offered a remarkable account of how Administration lawyers dealt with a series of legal questions in the weeks preceding Operation Neptune Spear, in which Navy SEAL Team Six killed Osama bin Laden in Abbottabad, Pakistan. The article is derived from an even more detailed account in Savage’s extraordinarily comprehensive Power Wars, which was published just this morning. Because the Times article has already generated a good deal of discussion and debate about the role of national security lawyering and the law in the Obama Administration — and the account in Power Wars is sure to do likewise — I thought perhaps it might be useful to home in upon this one discrete, albeit very important, subject of Savage’s book here at the outset of our Just Security symposium, before our contributors widen the lens to offer some broader perspectives on the book and what it might teach us about national security law and lawyering more generally.

Back in May 2011, just after bin Laden was killed, I wrote a long blog post in which I tried to piece together the Obama Administration’s views on a number of domestic and international law questions concerning the operation. In doing so, I necessarily was confined to relying upon various public statements that the President and other Administration representatives had made about Operation Neptune Spear, as well as other public pronouncements by the government describing its legal framework for the armed conflict with al Qaeda. Power Wars now offers a much more detailed account of how the Administration lawyers grappled with some of the most important legal questions raised by the proposals the President was considering for Abbottabad. The book confirms and clarifies several of the legal points I surmised about the bin Laden operation back in 2011. Moreover, it offers a very important case study in how the Administration tried to make good on the President’s commitment to act within applicable legal constraints in the use of force against al Qaeda.

The headline on the print edition of the Times story — “How 4 Lawyers Enabled Killing Of Bin Laden; Secret Memos Gave Legal Cover for Raid” — is deeply misleading. (The online edition headline is only slightly less tendentious: “How 4 Federal Lawyers Paved the Way to Kill Osama bin Laden.” Savage, of course, is not responsible for these headlines.) The purpose of the lawyering process Savage describes was not to “give cover” for the operation. It was, instead, simply to ensure that the operation was duly authorized and complied with a complex array of domestic and international laws — the everyday stuff of government lawyering, albeit in the context of an anything-but-everyday proposed course of action.

Many Americans — as well as commentators who insist that Presidents regularly blow past the law when it comes to the most important policy choices — might be surprised to learn that the President and his advisers devoted considerable attention to legal authorizations and constraints before going forward with the bin Laden operation. Yet Savage’s book and article depict a punctilious effort by Administration lawyers to identify, and deliberate upon, a series of legal questions, and to insist upon adherence to legal limits, even as to the planned attack upon the nation’s most notorious enemy. As I’ll discuss in this and a companion post, although some of the questions the lawyers discussed were fairly straightforward, as to a couple of others there was little precedent directly on point and thus the best answers were uncertain and contestable. Even so, the attorneys’ ultimate legal conclusions appear to have been sound, or at the very least reasonable and well-considered, as to each of the questions. 

Therefore, with one possible exception, I don’t see anything the least bit nefarious or improper about the process Savage describes. The lawyers simply did what attorneys ordinarily do: They made diligent efforts to to “work[] through every significant legal issue that they could identify” (p.262), and then advised their clients about the permissible scope, legal limitations, and conditions, of a proposed course of action — after which those clients proceeded (as far as we know) to adhere to the legal parameters set forth by the lawyers. That’s the system working as it should.

The possible exception, as others have noted, is Savage’s report (p.260) that National Security Advisor Tom Donilon chose not to inform the Attorney General (or others in the Department of Justice) of the planned Abbottabad operation until May 1, the day before the raid, after all the legal questions had already been asked and answered.

Jack Goldsmith writes that this exclusion of DOJ is “very hard to understand or justify.” I can certainly understand that reaction, especially because Savage’s book describes DOJ, and the Office of Legal Counsel in particular, as being so centrally involved in the analysis of so many other legal questions involving national security in the first two years of the Administration — at least through the time of Caroline Krass’s important opinion about the constitutionality of the initiation of force in Libya one month before the Abbottabad operation.

To fully understand and fairly assess Donilon’s decision, however, we’d need to know more than we do about his reasoning. What Savage’s account tells us is that Donilon was trying, quite understandably, to limit knowledge of the operation to the smallest possible group of officials — those with a strict “need to know.”   At first, according to Savage (pp. 257-59), only four attorneys were read in:  NSC Legal Adviser Mary DeRosa, CIA and DOD general counsels Stephen Preston and Jeh Johnson, and Jim Crawford, legal adviser to the Joint Chiefs of Staff.  Those four did the lion’s share of the legal analysis, according to Savage.  Later on, reportedly, White House Counsel Bob Bauer was included in the loop, as were two career attorneys enlisted to help–one working for DeRosa at NSC, the other for Preston at the CIA.  Savage quotes Donilon as remarking that it was “not clear why [AG] Holder’s input was necessary” in this case (p.260).

And perhaps it wasn’t so clear: After all, it is not at all unusual for DOJ, and the Office of Legal Counsel in particular, not to be consulted about legal questions concerning planned military operations, or every particular application of a covert action finding — or, for that matter, on 95% of the other legal questions that arise daily in the Executive branch — at least not unless there is interagency dispute or significant uncertainty about the answer. For example, DOJ has not typically addressed some of the sorts of jus in bello questions at issue in the bin Laden case (e.g., involving denial of quarter and proper burial), while the analysis of others (such as whether prior notification of the oversight committees was required) was quite straightforward, and presumably informed by earlier OLC advice.

On the other hand, as I discuss below, in this case the question involving Pakistani sovereignty and the U.N. Charter was tricky and very important, and its answer would likely set a significant precedent. So perhaps the better and more appropriate course of action would have been to bring DOJ/OLC into the loop on at least that question, especially once the operation had been briefed beyond the original core circle to include, e.g., the Secretary of State and the White House Counsel. (Having said that, I should add that I doubt Donilon excluded DOJ because he was afraid that the AG or OLC would offer especially unwelcome legal advice: There’s no apparent reason to think that those attorneys would have materially disagreed with the analysis that the other national security lawyers developed on the questions relating to the bin Laden operation.)

Moving on to the substance, in the remainder of this post and a companion post I’ll try to unpack and analyze in greater detail the substance of the legal questions that Savage identifies as having been the principal subjects of the Administration attorneys’ deliberations and advice. I will assume, for purposes of these posts, that Savage’s account is accurate (which it almost certainly is in broad outline and in most respects, even if not in every detail).*

In this first post, I’ll focus upon what surely was the most important and difficult question the lawyers confronted — namely, whether going forward with Operation Neptune Spear, without Pakistani consent, would constitute a use of force against the territorial integrity of Pakistan that Article 2(4) of the United Nations Charter prohibits.  [UPDATE:  Second post is here.]

The UN Charter and Pakistani Sovereignty

Article 2(4) of the Charter generally requires a state to “refrain in [its] international relations from the . . . use of force against the territorial integrity or political independence of any state.” This prohibition does not apply if the force is authorized by the United Nations Security Council or the “host” state consents to it.

The bin Laden operation was a “use of force,” and it was not authorized by the U.N. Security Council. Nor, as Savage confirms, did Pakistan consent to it. Therefore it would have violated Article 2(4) . . . unless it was a legitimate act of self-defense, for Article 51 of the Charter provides that “[n]othing in the present Charter shall impair the inherent right of individual or collective self-defence if an armed attack occurs against a Member.”

Whether the bin Laden operation would have been a proper exercise of the “inherent” right of self-defense was, as Savage notes, a question of international law — but, importantly, it wasn’t only that. The UN Charter is a multilateral treaty to which the United States is a party. Therefore the Charter is domestic law, too and, as I explain below, the President has a constitutional obligation not to place the U.S. in breach of Article 2(4) unless Congress has authorized him to do so.

al Qaeda’s armed attacks on the U.S., including on 9/11, triggered the U.S.’s right to use force in self-defense against al-Qaeda in order to prevent further such attacks by that armed group, as the Security Council recognized. Even so, in order to exercise the “inherent” right of self-defense against a nonstate actor in the territory of a state that has not consented to it, the acting state (here, the United States) must satisfy the customary law norms of necessity and proportionality, which are well-recognized conditions for exercising the “inherent” right of self-defense.

The bin Laden operation certainly appears to have been proportionate to the continuing threat: The U.S. had concluded that bin Laden was continuing “to plot attacks against the United States,” and the operation was deliberately designed not to result in significant casualties to Pakistani civilians or civilian property.

But did it satisfy the requirement of “necessity”? That was the difficult question confronting the lawyers. If Pakistan would have consented to the operation, or if Pakistan itself would have captured or killed bin Laden at the U.S.’s request, then the U.S.’s unconsented use of force would not have been necessary. This is often referred to as the “unwilling or unable” test, which is best understood as an application of the jus ad bellum requirement of necessity: In short, when State A considers the use of force against a nonstate actor in order to stop the threat of future armed attacks from that actor, and the nonstate group has taken safe haven in State B, which has not consented to State A’s use of force, the requirement of necessity is satisfied only if State B is not itself willing and able to address the threat to State A.

And that’s where the legal/empirical problem got very tricky in the bin Laden case, because (as Savage reports) U.S. officials could not determine for certain whether or not Pakistan would have consented to the U.S. operation, or was itself willing and able to eliminate the bin Laden threat. Why not? Because U.S. officials concluded that even to ask Pakistan such questions might well itself undermine the prospect of addressing the threat: According to Savage, “the administration feared that the Pakistani intelligence service might have sanctioned Bin Laden’s presence; if so, the reasoning went, asking for Pakistan’s help might enable his escape.” In other words, if the U.S. were to ask Pakistan for its consent and/or assistance, and Pakistan said “no,” that answer would trigger the United States’s right to act unilaterally . . . but that same inquiry and negative response might also make it virtually impossible for the United States to successfully act in self-defense to stop the threat posed by bin Laden, either because the ISI would alert him to the proposed action, or because Pakistan itself might gear up to have its forces interdict the U.S. forces.

Of course, if the “fears” of the U.S. officials had been unreasonable — if there were, in fact, little risk that the ISI would tip off bin Laden or that Pakistan would do anything else to undermine the operation — then presumably it would not have been permissible for the United States simply to assume that Pakistan would not consent or would be unable and unwilling to stop the threat.

But let’s assume (as the lawyers apparently did) that those fears were well-founded — for instance, that if the United States gave Pakistani officials a head’s up about bin Laden and Abbottabad, there was a good chance the ISI would have tipped off bin Laden. (Savage cites the fact that bin Laden had been living for years in a suspicious compound in the heart of the Pakistani security establishment as “circumstantial evidence that ISI elements might have been helping him or at least knew he was there” (p.260).)  In that case, wouldn’t it have been reasonable for the lawyers to conclude that the United States’ use of force without prior Pakistani notification/coordination was, more likely than not, necessary to interdict the threat posed by bin Laden?  I read Savage’s account to suggest that the lawyers concluded it would be. Was that correct?

To be sure, there’s very little extant custom, one way or the other, that speaks directly to this particular difficulty in implementing the “unable or unwilling” application of the necessity requirement for self-defense, in cases where there is a good chance officials within the host state would undermine the effort.  (That’s why Ashley Deeks proposes that states ought to develop a future customary “test” of the sort I’ve described above.) But since such obstruction would itself entirely defeat any possibility of being able to act in self-defense, I can’t imagine any state taking the risk of seeking consent or cooperation under such circumstances — which probably explains why (to my knowledge) no states other than Pakistan itself have accused the U.S. of having violated Article 2(4) of the Charter in the bin Laden operation.

Accordingly, although perhaps the customary law of self-defense does not yet conclusively establish an answer, I think it was certainly reasonable for the lawyers to conclude that the “inherent” right of self-defense, preserved in Article 5 of the Charter, did not require the United States to ask Pakistan for its consent and/or cooperation before the fact.

Was Breaching the Charter a “Trump Card”?

Savage reports that “[t]here was also a trump card” (p.264). “While the lawyers believed that Mr. Obama was bound to obey domestic law,” he writes, they also believed he could decide to violate international law when authorizing a ‘covert’ action, officials said.” The suggestion here is that at least some of the lawyers thought that even if the bin Laden operation would violate Article 2(4) of the UN Charter, the President could order it anyway as a matter of domestic law, even though it would obviously violate the U.S.’s international law obligation to comply with Article 2(4) of the Charter.**

It’s not necessary here to wade too deeply into this question, since Savage does not suggest that the lawyers actually advised the President that he was authorized to put the United States in breach of Article 2(4). In other words, even if they contemplated a “trump card,” there’s no indication that the lawyers played it. And for good reason: It’s highly unlikely that this President would decide, and assert the authority, to breach Article 2(4) of the Charter, even if that were a legally available option. So I’ll offer only a few brief thoughts on the question here, distinguishing between two often-invoked versions of the argument.

One version of the so-called “trump card” argument is that Article 2(4) is not a “self-executing” treaty provision, and that therefore, although the provision is of course binding on the United States as a matter of international law (namely, the IL obligation to abide by one’s treaty commitments), the President has no obligation under domestic law to abide by it–that is to say, it is not covered by his constitutional obligation to “take care that the laws be faithfully executed,” because that constitutional obligation only extends to domestic law, and a non-self-executing provision is (according to this argument) not domestic law.***

This argument appears in a 1989 opinion of the Office of Legal Counsel with respect to Article 2(4) itself.  OLC reasoned  in that opinion that because Article 2(4) of the Charter is non-self-executing–in the sense that “it was not intended by the United States at the time of ratification to affect the criminal jurisdiction of federal courts,”4B Op. O.L.C. at 548, unless and until Congress provides otherwise–therefore it is “not legally binding on the political branches,” and thus “as a matter of domestic law, the Executive has the power to authorize actions inconsistent with Article 2(4) of the U.N. Charter.”

Quite frankly, this was a non-sequitur in 1989, and it remains so.  The fact that a court cannot enforce a treaty provision without further legislative action — because the President and the Senate (arguably) determined that it would be “non-self-executing” in that sense — does not say anything about whether the provision imposes a binding norm under domestic law for executive conduct.  That is to say (and to invoke the seminal Court case on this subject), the fact that Congress “must execute the [treaty provision] before it can become a rule for the Court,” Foster v. Neilson (Marshall, C.J.), does not mean that Congress must pass a law to implement the treaty provision before it can become a rule for the Executive.  The two are distinct questions, and they don’t necessarily share the same answers.  [For more along this line, see Deborah Pearlstein.]

Unfortunately, some dicta in the Supreme Court’s decision in Medellín v. Texas appeared to offer support for the argument:  Non-self-executing treaties, wrote the Chief Justice, “are not domestic law unless Congress has . . . enacted implementing statutes.”  [UPDATE:  Such an “is not domestic law” statement might make sense with respect to a provision that can only be implemented with Congress’s involvement–such as a provision requiring the United States to impose criminal sanctions on certain conduct.  (The President can’t enforce such a prohibition unilaterally, without a criminal statute.)  But where, as here, the treaty obligation is prohibitory–where it requires states not to do X (“All Members shall refrain in their international relations from the threat or use of force . . . )–it would be a rare case, indeed, in which the Senate and President that negotiated the ratified the treaty would intend that domestic law would permit the Executive to breach the nation’s solemn obligation not to violate that prohibition.]

Fortunately, six months after Medellín, the Senate Foreign Relations Committee, responding to the confusion sowed by the Supreme Court, clarified that even when the Senate expressly designates a treaty provision as “non-self-executing,” that term should not be construed to excuse the President of his responsibility under domestic (as well as international) law to comply with such a provision:  “The committee believes it is of great importance that the United States complies with the treaty obligations it undertakes.  In accordance with the Constitution, all treaties — whether self-executing or not —are the supreme law of the land, and the President shall take care that they be faithfully executed.”

Article 2(4) of the U.N. Charter plainly does impose a prohibition that was designed, and expected, to be binding on the Executive even if not judicially enforceable.  Indeed, it’s a fairly preposterous notion, I think, to suggest that when the U.S. ratified the Charter, President Truman and the 1949 Senate would have seriously entertained the notion — let alone intended — that Article 2(4) would not be binding on the President as a matter of U.S. law.

[UPDATE:  It’s important to note, in this respect, that the terms “self-executing” and “non-self-executing” are mere labels that one may use as a shorthand to describe the legal effects of particular treaty provisions.  Sometimes the President and the Senate append such a label to ratification documents during the treaty-making process, in which case the label means whatever those actors intended it to mean for purposes of domestic U.S. law.  As to those Senate-designated provisions, the labels themselves don’t have any inherent or invariable meaning, unless the political actors decide, over time, that they ought to become terms of art–terms that should be uniformly understood as always having some effects but not others (which is what the SFRC is trying to do in the Report quoted above, i.e., laying down a marker of what it intends to convey when it uses the term “non-self-executing” to describe treaty provisions).

Most treaties, however, including the Charter, do not bear a label one way or the other, even in their internal U.S. ratification documents.  In such cases, then, for someone to characterize a provision as “self-executing” or “non-self-executing” does not itself magically afford that provision certain legal properties.  Those terms are not some sorts of one-size-fits-all on/off switches that have uniform legal effects merely because some actor–a President, a Court, an academic–invokes them.  They are, instead, a shorthand for describing one or more legal effects of the treaty provision.  And therefore, to determine what those effects are, one can’t simply “decide,” in any abstract sense, that a provision is, or is not, “self-executing.”  That is to say, at least with respect to a provision that the President and the Senate have not designated one way or the other, it is putting the cart before the horse to decide in the first instance whether a provision is self-executing and then to use that determination in order to ascertain its legal effects.  One must instead inquiry directly what the legal effects are–at which point it might be useful to append a label (although, quite frankly, I think the labels will continue to be more obfuscating than clarifying unless and until all of the relevant actors mean the same things by the use of the particular adjectives).

And so, with respect to Article 2(4) of the Charter, in 1980 the Office of Legal Counsel determined (I assume correctly) that “it was not intended by the United States at the time of ratification to affect the criminal jurisdiction of federal courts.”  4B Op. O.L.C. at 548.  OLC then labeled Article 2(4) as “non-self-executing” by virtue of that determination about its effect on court jurisdiction.  That label, however, does not then magically trigger other legal effects, such as affording the Executive a license under domestic law to put the United States in breach of the provision or, as the Court in Medellin put it, rendering the provision “not domestic law unless Congress has . . . enacted implementing statutes.”

Here’s another way at getting at the same point:  If you are someone who–in contrast to the Senate Foreign Relations Committee–insists upon using the term “non-self-executing” to refer to a treaty provision that is not domestic law unless Congress has enacted implementing legislation, well, in that case . . . Article 2(4) is not non-self-executing (that is to say, it is self-executing), given your use of those terms.]

There is, however, another possible “trump card” argument, which is a bit trickier — namely, that Congress has actually authorized the President to breach U.S. treaty obligations, as long as he does so by way of covert action. (Congress has the constitutional authority to enact legislation authorizing the U.S. to breach a treaty, although there’s a general presumption that it does not intend to do so.) This argument depends upon the language of the covert action provision of the National Security Act, 50 U.S.C. 3093(a)(5), which provides that a presidential finding authorizing covert action “may not authorize any action that would violate the Constitution or any statute of the United States.” Some have suggested that Congress’s deliberate decision not to refer to “treaties” in that prohibition should be understood as an implicit authorization to the President to use covert action — defined in the statute as “an activity or activities of the United States Government to influence political, economic, or military conditions abroad, where it is intended that the role of the United States Government will not be apparent or acknowledged publicly,” 50 U.S.C. 3093(e) — in a way that would place the U.S. in breach of its treaty obligations, including even an obligation as important as the injunction of Article 2(4) that a state shall “refrain in [its] international relations from the . . . use of force against the territorial integrity or political independence of any state.”

The hint of such an argument appears in Caroline Krass’s written submission to the Senate in 2013, in connection with her confirmation hearing to become CIA General Counsel. Krass wrote that “[b]y this language [in 3093(a)(5)], Congress did not prohibit the President from authorizing a covert action that would violate a non-self-executing treaty or customary international law. Such a covert action would, however, need to comply with treaty provisions implemented in a domestic statute.” As Ashley Deeks has noted, this response “strongly implies that the President may authorize covert action that violates non-self-executing treaties . . . , though [Krass] does not state this explicitly.” I think Deeks is correct to say that the Krass answer “suggests that insofar as Article II authorizes the President to violate . . . treaties, the Act does nothing to restrict such action” — but in that case, we’d simply be back to the question of whether the President has the constitutional authority to breach Article 2(4), and as I’ve noted above, I don’t think he does.

There is reason to be skeptical of the argument that section 3093(a)(5) itself authorizes a breach of Article 2(4), not least because of the Charming Betsy canon that statutes should be construed on the presumption that Congress does not intend to authorize the U.S. to violate international law. Moreover, such a conclusion would be in at least some tension with the fact, accepted and endorsed by the Obama Administration, that the 2001 AUMF requires the Executive to act in compliance with international law in prosecuting the armed conflict against al Qaeda: Could it really be the case that the covert action statute authorizes the use of lethal force in that same conflict in a manner the AUMF would prohibit?

In order to assess the question fully and fairly, however, one would need to know much more about the genesis of section 3093(a)(5), its historical application and understandings in the political branches (e.g., by the intelligence committees that oversee covert actions; see also my discussion of a related provision of the statute, section 3231), and its relationship to the AUMF.

Thankfully, such a deep dive into the intricacies of the covert action statute is not necessary here because, as noted above, there’s no reason to think the President decided to breach Article 2(4), or even that the lawyers advised him that he had such an option, in Pakistan in May 2011.

In my companion post I’ll address the jus in bello questions Savage identifies the attorneys as having considered (including the requirement to accept an offer of surrender; burial at sea; and proportionality), as well as the question of congressional notification.


* Even a reporter as thorough and conscientious as Savage — and it’s hard to imagine there’s anyone more thorough and conscientious, at least on these issues — is necessarily dependent upon his sources’ accounts, which are often biased, incomplete and contradictory. Indeed, Savage himself admirably acknowledges that he has updated and clarified certain details for the e-book and subsequent printings of Power Wars as he obtained further sourcing. Therefore it’s important for readers to keep in mind that even in the most thorough and accurate national security reportage, not every detail can be taken for granted. Having said that, I should emphasize that I have no reason to think any of the reported details I discuss in this post are inaccurate. (I left my government position before the events in question occurred and therefore have no direct personal knowledge of what transpired.)

** This is not the same as the famously contested “Paquete Habana” question of whether and in what circumstances the President can violate customary international law. Whatever the answer to that question might be, treaties are the supreme law of the land, and the President has an obligation to take care that such treaties are faithfully executed.

*** [UPDATED] The minor premise of this argument–that the President’s Take Care obligation does not extend to purely international law–is, of course, itself highly contested.  But I’m assuming for present purposes that it’s correct.