Ryan posted last week about varying accounts of who was killed in an attack on an apparent wedding party in rural Yemen on December 12, 2013.  A Human Rights Watch report quotes witnesses and family members as saying that some of the victims were civilians.  Yemeni government sources offered several different accounts of who was killed.

There hasn’t been any official, public U.S. account or assessment of the strike.  Indeed, as far as I know the U.S. has not officially acknowledged U.S. involvement in the strike at all.  (The widespread assumption that the U.S. was responsible appears to be based primarily on the fact that parts of Hellfire missiles were recovered from the site.  Yemeni officials reportedly did provide some apologies and compensation to at least some of the families of the victims; but I’m not aware that Yemeni officials have said anything about which states were responsible.)

An Associated Press story published last Thursday, however, had this to say about the internal U.S. assessment of the strike:

Three U.S. officials said the U.S. government did investigate the strike against [Shawqi Ali Ahmad] al-Badani[, a midlevel al-Qaida leader, said to have been the target of the attack]– twice–and concluded that only members of al-Qaida were killed in the three vehicles that were hit.

The officials said the Pentagon can’t release details because both the U.S. military and the CIA fly drones over Yemen.  By statute, the military strikes can be acknowledged, but the CIA operations cannotThe officials said that if they explain one strike but not another, they are revealing by default which ones are being carried out by the CIA.

The purpose of my post here is to explain that this account of why the Pentagon “can’t release details” of its investigation makes little sense, and unfortunately leaves a deeply misleading–that is to say, inaccurate–account of the law and the reasons that the U.S., like other nations, sometimes refuses to acknowledge its involvement in particular actions overseas.  Unfortunately, such misunderstandings are very common, in ordinary reporting and in public debates.  Therefore I thought it would be worth unpacking the issue a bit.

Before I proceed to the substance, however, it’s worth noting that there’s good reason to be skeptical of the accuracy of the A.P. account.  The reporter’s reference to “three U.S. officials”–rather than, say, “high-level administration officials”–is a commonly understood signal that at least one, perhaps all three, of the “officials” work in Congress, not the Executive branch.  The sources themselves, therefore, may not have been involved in the reported investigations or any legal or policy assessments concerning secrecy and nonacknowledgement.  (In other words, they may not know what they’re talking about.)  Moreover, the key misleading sentence, about the law–”By statute, the military strikes can be acknowledged, but the CIA operations cannot.”–is not directly attributed to “officials” at all; therefore it might be a paraphrase of something someone said, at best . . . or perhaps reflects only the reporter’s own uncritical assumption of the law.

All of which is to say:  The A.P. story is hardly an authoritative account.  Now, on to the substance:

I.  Secrecy v. Nonacknowledgement

For starters, it’s important to distinguish between secrecy, on the one hand, and nonacknowledgement, on the other.  There are many reasons for the United States, or another state acting outside its borders, to wish to keep matters secret–protection of sources and methods being among the most common and most important.  In other cases, a state will have reason not to officially acknowledge its involvement in a certain action, even if the action is no longer secret in any real sense, i.e., where it has been widely reported and where it is widely assumed (and perhaps even unofficially leaked) that the state in question was involved.  What could those reasons for nonacknowledgement be?

Most commonly, as Jack Goldsmith has noted, a refusal to officially acknowledge U.S. involvement may be thought necessary “because other nations participating in the support insist on USG deniability as a condition of participation,” or because, in cases where the host state is not consenting, “explicit acknowledgment of USG [involvement] would raise the stakes high enough . . . that [such nations] would feel compelled to retaliate more openly and aggressively.”  In other words, official acknowledgement can cause distinct diplomatic and foreign-relations harms, even if the facts in question are widely reported or rumored in the public domain.  These are, in any event, the justifications for nonacknowledgement that the U.S. often invokes in litigation challenging such refusal to acknowledge.  See, for example, the cases and rationales discussed in footnotes 1-5 of Judge Katzmann’s opinion concurring in the judgment in Wilson v. CIA, 586 F.3d 171, 197-99 (CTA2 2009); see also the majority opinion in Wilson at 186 (“As a practical matter, foreign governments can often ignore unofficial disclosures of CIA activities that might be viewed as embarrassing or harmful to their interests.  They cannot, however, so easily cast a blind eye on official disclosures made by the CIA itself, and they may, in fact, feel compelled to retaliate.”).

II.  The Law

Keeping in mind the distinction between secrecy and nonacknowledgement, let’s turn now to the law.  (I apologize in advance that this is somewhat detailed.  Please bear with me.)

The A.P. account of the Yemen strike asserts that U.S. law requires nonacknowledgement when the CIA is involved.  That is wrong.  There is no statute, or any other domestic law, that provides that “military strikes can be acknowledged, but . . . CIA operations cannot.”

Indeed, it is not the law that determines whether the U.S. does or does not acknowledge strikes, or any other actions overseas.  U.S. law generally does not require or prohibit nonacknowledgement at all.  Rather, as explained above, nonacknowledgement is typically a function of diplomatic and foreign policy considerations.

a.  The Covert Action Statute.

Now, as it happens, U.S. law does have something to say about nonacknowledgment, but it’s very different from what is stated in the A.P. story:  A provision of the National Security Act, 50 U.S.C. section 3093 [formerly 413b(e)], generally defines as “covert action” any “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.”  The statute then includes certain exceptions–actions that are not deemed “covert,” even if they are designed to influence political, economic, or military conditions abroad, and even if they are undertaken with the intent that the role of the U.S. Government will not be apparent or acknowledged publicly.  The most important of these exceptions, which I’ll discuss below, is for “traditional . . . military activities or routine support to such activities.”

The primary function of this statutory definition is restrictive–it describes those U.S. actions that are subject to the particular limitations and conditions prescribed elsewhere in section 3093.  There are two major conditions for activities that fall within the definition:  (i) they must be authorized, almost always in advance (except in certain emergency situations), by a written presidential finding in which the President “determines such an action is necessary to support identifiable foreign policy objectives of the United States and is important to the national security of the United States,” and in which the President specifies which departments, agencies or entities, inside or outside the U.S. government, may fund or otherwise participate in any significant way in such action; and (ii) such actions, and the presidential findings, are subject to detailed and extensive requirements of reporting to the congressional intelligence committees.  The first condition ensures that the President is accountable for covert actions (it eliminated the problem of “plausible presidential deniability”), and, in effect, provides for extensive oversight by the National Security Council.  The second condition, of course, is designed to ensure comprehensive and effective legislative oversight.  (There are also two substantive limitations:  an action that falls within the covert action definition (i) may not “violate the Constitution or any statute of the United States”; and (ii) may not be “intended to influence United States political processes, public opinion, policies, or media.”)

What follows from this understanding of section 3093 and the definition of “covert action”?

First, and most importantly, the primary function of the statute is to prescribe what must happen if an activity is intended to be unacknowledged, rather than prescribing whether an activity can or should be acknowledged in the first instance.  And this brings us to a common cart-and-horse misunderstanding:

It is often assumed that a particular activity must remain unacknowledged because the President has signed a covert action finding authorizing that activity.  But that gets things backwards.  Rather, the President must sign a covert action finding if–that is, because–he has already determined that U.S. involvement in an activity to influence political, economic, or military conditions abroad should or must be unacknowledged–a determination that, as noted above, is a function of diplomacy and policy, not domestic law.

Second, with one primary exception (discussed below), the legal rules themselves do not distinguish between agencies.  The covert action statute applies, by its terms, to all “departments, agencies, or entities of the United States Government.”  Thus, if an action falls within the covert action definition, it is subject to the rules of the statute–it must be subject to a presidential finding and reported to the intelligence committees–regardless of which agencies are involved.

More generally, as far as the law is concerned, both the CIA and DOD can engage in activities that are not officially acknowledged; and likewise, both can engage in overseas activities that are officially acknowledged, as long as they have an independent source of affirmative authority for doing so (such as the AUMF, the “fifth function” provision of the National Security Act (see below), or Article II).  Moreover, and as section 3093 indicates, the key question is typically not whether the involvement of any particular agency should be officially acknowledged, but instead about whether the United States’s involvement should be acknowledged.  (It is possible, I suppose, that in a particular case there might be reason to acknowledge the actions of one agency but not another–for instance, if that is what an ally insists upon as a condition of cooperation.  But again, that would be a function of diplomatic particulars, not the law.)

The principal exception to the parallel legal treatment between DOD and the CIA is that “traditional military activities” are excluded from the definition of covert action.  What follows from this exception?  Not that the military cannot engage in unacknowledged activities — it can — but instead that the military in some cases may be exempt from the detailed section 3093 finding/oversight/reporting regime.  If the armed forces engage in a “traditional” military activity, then by definition that activity doesn’t fall within section 3093, and therefore it need not be authorized by a presidential finding, and it need not be reported to the intelligence committees, even if that activity is “designed to influence political, economic, or military conditions abroad” and is undertaken “with the intent that the role of the U.S. Government will not be apparent or acknowledged publicly.”  (Of course, DOD actions are subject to oversight by the armed services committees, and can be subject to internal National Security Council oversight or coordination if the President directs it.  But these constraints upon the military’s unacknowledged activities are as a general rule less robust than the requirements of section 3093.)  The question of which DOD activities are “traditional” is beyond the scope of this post, except to note that unacknowledged DOD actions can be “traditional,” at least as long as they are undertaken “by military personnel under the direction and control of a United States military commander . . . preceding hostilities which are anticipated . . . involving U.S. military forces, or where such hostilities are ongoing, where the fact of the U.S. role in the overall operation is apparent or to be acknowledged publicly.”  S. Rep. No. 102-85, at 46 (1991).

b.  The permissibility and impact of acknowledging previously unacknowledged U.S. involvement.

So that’s the basic, relevant legal framework.  But before I move on to the Yemen strike, it’s important to address one more issue, one of timing.  (If press accounts are accurate, this scenario might roughly describe what has occurred in Yemen.)  Let’s say that at Time A, the President decides that a particular U.S. activity (say, some sort of use of force in a particular nation) must remain unacknowledged; accordingly, he signs a finding authorizing one or more agencies to undertake that activity as a covert action, and the agencies so authorized begin to engage in unacknowledged uses of force in the nation in question.  Then, at Time B, the President decides that U.S. involvement in that nation can now be acknowledged.  Subsequent to Time B, does it remain necessary not to acknowledge the actions of the agencies that have been acting pursuant to the earlier covert action finding–a finding that was predicated upon an intent not to acknowledge U.S. involvement?  I don’t think so.  Once the President has determined that nonacknowledgement of a particular activity is no longer necessary, acknowledgement of the U.S. role in that activity becomes permissible, even as to those actions that are being undertaken pursuant to a covert action finding.

The most widely known example of this was Operation Neptune Spear, the operation to capture or kill Osama bin Laden.  In a PBS interview with Jim Lehrer shortly after the operation, then-CIA Director Panetta stated that “this was what’s called a ‘title 50’ operation, which is a covert operation, and it comes directly from the president of the United States who made the decision to conduct this operation in a covert way.”  Assuming this is true, it means that the President originally intended that U.S. involvement in the operation would not be officially acknowledged, at least in the event the operation went awry.  But of course, mere hours after the operation, the President did acknowledge U.S. involvement.  At that point, there was no legal bar to discussing the CIA’s involvement . . . hence, the Panetta interview.

The same is true with respect to ongoing activities.  Once the President decides that an activity can be acknowledged, then . . . it can be acknowledged, even if some of the agencies involved had previously been acting, in whole or in part, pursuant to a covert action finding.  (This is not to say that the covert action finding itself can or should then be acknowledged in toto–it might be the case, for instance, that the finding authorizes more than one covert activity.)

The fact that acknowledgment of a covert action is lawful, however, does not necessarily mean that such acknowledgement would have no legal effect whatsoever.  The most acute additional question is this:  Would formal acknowledgment of a CIA covert action thereby eliminate any affirmative authority the CIA has to continue the activity?  In a recent article, Bobby Chesney reports [see note 214] that he has “been told by a number of sources” that whether the CIA has authority to engage in overt, or acknowledged, activities, “has been the subject of intense disagreement within the government, notwithstanding that the language of the so-called ‘Fifth Function’ in the National Security Act [see 50 U.S.C. 3036(d)(4)] does not appear to require that an operation be covert in order to come within the CIA’s competencies.”

I do not think it is likely that official acknowledgment of U.S. responsibility would eliminate affirmative authority for the CIA to continue what had previously been an unacknowledged activity.  There are at least five possible sources of CIA authority (depending on the activity in question):  the “Fifth Function” provision of the National Security Act; article II of the Constitution; a force authorization statute such as the AUMF; appropriations/authorizations laws; and section 3093 itself.  (This is not the occasion to debate the existence or scope of these authorities.  I am merely identifying the sources that are invoked from time to time.)  With the possible exception of appropriations laws — which may incorporate by reference certain agreements that are not public, and which we therefore have no way of evaluating — none of these potential sources of authority would appear to require continuing nonacknowledgement as a condition of CIA activity.  (Even section 3093 itself, to the extent it were seen as affirmative authority for an agency activity, might permissibly be construed only to require an intent not to acknowledge U.S. involvement at the time the President signs the finding, but not to cut off all affirmative authority if the President later deems it appropriate to acknowledge U.S. involvement.)  [UPDATE:  My apologies for not noticing and flagging this sooner:  In her recent submitted Q&As for her nomination to be CIA General Counsel, my former colleague Caroline Krass was asked the question directly, and she answered unequivocally (see page 6) that, pursuant to the “fifth function” statute, the President or the DNI could direct the CIA to engage in activities to influence political, economic, or military conditions abroad where it is intended that the role of the United States Government will be apparent or acknowledged publicly.]

A recent event might be a good test of this question.  In a Senate Foreign Relations Committee hearing last September, in response to a question from Senator Corker about why U.S. aid to the Syrian rebels had been delayed, Secretary of Defense Hagel stated that in “June of this year . . . the president made a decision to support lethal assistance to the opposition, as you all know,” and that “[t]his is, as you know, a covert action.”  (So much for official nonacknowledgement of U.S. involvement.)  Did this public testimony mean the end of the assistance program?  I don’t know . . . but I doubt it.

In any event–and this is the important point for present purposes–the only context in which that question of continuing affirmative authority for CIA activity would even arise is where there is official acknowledgement of the U.S. involvement in the activity; surely, any source of CIA authority would not lapse by virtue of the sort of acknowledgement-by-negative-implication to which the A.P. story refers.  This potential issue of continuing authority, then, even if it is more acute than I have suggested above, is not a reason not to acknowledge any DOD activities.

III.  Back to the Yemen Strike (and its Investigation)

So, what does all this tell us about the Yemen strike and any U.S. investigation of it?

a. The Strike.

Before we address the investigation, it appears that the U.S. has not even officially acknowledged whether it was or was not involved in the strike.  We don’t know why that is so.  The President has officially reported “direct action” by the U.S. against AQAP in Yemen, so there is no general policy of not acknowledging U.S. use of force in that country.  It is possible there is some other reason not to acknowledge specific strikes in Yemen.  But it’s not because of any legal requirement of nonacknowledgement.

And it should not be because of the reason offered in the A.P. account, namely, that “if they explain one strike but not another, they are revealing by default which ones are being carried out by the CIA.”  That explanation doesn’t make much sense, for several reasons.

First, it’s not logical even on its own terms — acknowledgment of one or more strikes would not tell anyone, one way or the other, about which agencies are or are not involved in strikes that are not acknowledged.

Second, as noted above, the relevant diplomatic consideration (and the criterion for the covert action definition, too) is generally whether the involvement of the United States will or will not be acknowledged–not whether the involvement of one U.S. agency or another is acknowledged.

Third, and most fundamentally, the whole point of official nonacknowledgement, at least in the case of an activity that is already transparent as a general matter, is not so much to preserve secrecy, or to prevent observers from inferring U.S. involvement (let alone involvement of a particular agency), but instead to avoid official acknowledgement of U.S. involvement, even with respect to information that is already in the public domain, because such official acknowledgment can cause distinct harms.

Thus, if Operation A is officially acknowledged, and Operation B is not, the distinct values of nonacknowledgement of Operation B remain intact, no matter what may be inferred from acknowledgement of Operation A.  (On this general point, see pages 198-201 of Judge Katzmann’s concurrence in Wilson, pp. 52-55 of the slip opinion.)

b.  The Investigation.

Finally, why is the Executive apparently unwilling (so far, anyway) to reveal the details of its own investigations of the Yemen strike?

Well, if the U.S. was involved in the strike and has determined, for whatever reason, that such involvement must remain officially unacknowledged, then of course it would not be able to release the investigation reports for that reason alone.

Alternatively, even if the U.S. were prepared to acknowledge involvement in the operation, and to disclose the conclusions it has reached in its investigations, it’s possible that some of the evidence on which it relied, both in deciding to use force and in investigating the strike, was derived from confidential sources that cannot be revealed.  If so, the Executive might be at work now to figure out how to publicly justify its use of force and/or its investigatory conclusions without burning such sources.

All of which is to say that we don’t yet know the reasons for the U.S.’s silence . . . but they are almost certainly not what the A.P. has reported.