Should We Rely on Congress?
Oversight and the Problem of Executive Privilege
We have now seen why we should be cautious about relying upon judges to determine whether the executive is employing secrecy responsibly. Does it make sense to rely on Congress instead? As we noted in chapter 1, Congress’s record on this count is less than promising. Though the Framers claimed that the separation of powers would allow Congress to oversee the executive’s conduct of war and diplomacy, in practice lawmakers have invariably found themselves in the dark in precisely those instances where they ought to have been best informed—that is, in cases where presidents have made far-reaching national security decisions under the veil of secrecy.
Why has Congress struggled to stay abreast of the president’s secret actions and policies? The blame is usually placed on the fact that Congress relies on the executive to keep it properly informed.1 Congress “remains at the forbearance of the executive in terms of the intelligence it is given,” Britt Snider has written, because it “cannot request information it does not know exists.”2 To be sure, the establishment of intelligence oversight committees in Congress has led to expanded oversight of routine intelligence community activity.3 For instance, the numbers of briefings and testimonies provided by the intelligence community have escalated sharply since the mid-1970s.4 “Congress now receives nearly every intelligence item the executive does,” according to Gregory Treverton.5 But these developments notwithstanding, the record shows that Congress tends to become aware of far-reaching national security initiatives rather late in the day, the best-known examples being the Iran-Contra affair and, more recently, the politicization of intelligence on Iraq’s weapons program, the establishment of secret prisons, and the use of “enhanced” interrogation techniques and warrantless surveillance.6
This record may seem puzzling. Why, one might wonder, does Congress not demand fuller access to national security information, especially about operations that have the potential to violate domestic and international law? There are a number of explanations on offer. One focuses on the overseers themselves, a great many of whom, Loch Johnson has argued, have acted not as watchful “guardians” of the public trust, but rather as uncritical “cheerleaders” for the intelligence community.7 Another explanation draws attention to the high turnover and large memberships of the intelligence oversight committees, which, according to two recent congressional reports, gives members little chance or incentive to develop expertise in intelligence affairs.8 A third explanation emphasizes the relative obscurity of intelligence matters, which, Frank Smist has argued, offer little opportunity for patronage or grandstanding and therefore do not generate much interest among members of Congress.9
These accounts undoubtedly help us understand why Congress sometimes makes little or no effort to conduct oversight. But is a lack of will the only problem? The explanations cited above assume that Congress can have access to national security information—so long as it really wants it. Is the story really that simple? At least some scholars seem to think so. Mark Rozell, Gary Schmitt, David Crockett, and Neal Devins, among others, have suggested that we need “do nothing” about the fact that Congress depends on the president for information, because Congress can use its wide-ranging powers to compel the president to part with the information it needs to perform oversight.10 If Congress “wants something badly enough,” Crockett has argued, “it will use its political resources to full effect.”11
Consider, then, the powers that Congress enjoys. Foremost among these is the power to legislate. In principle, Congress can use this power to limit the kinds of activities that an unforthcoming president can undertake in secret. This approach draws on the concept of “generalization”: the idea that “if a particular decision cannot be disclosed in advance, the general type of decision can be discussed publicly, its justifiability in various hypothetical circumstances considered, and guidelines for making it in those circumstances formulated.”12 The appeal of a generalized restriction lies in the fact that it appears to eliminate the sorts of moral and legal ambiguities that officials often utilize to justify questionable behavior that is shrouded in secrecy. Because it leaves no doubt about the sorts of covert actions or secret policies that lawmakers are willing to condone, a statutory restriction or prohibition affords officials little “wiggle room” (and little excuse once violations are discovered).
However, the closer we come to practice, the less useful a generalized restriction becomes. The problem, as Dennis Thompson has noted, is that the complex and unpredictable nature of political life can make it difficult for lawmakers to specify restrictions in advance. Consider, for example, the recent controversy surrounding the NSA’s warrantless surveillance program. When critics charged that the program violated the law, the Bush administration defended the program by pointing to Congress’s Joint Resolution on the Authorization for Use of Military Force Against Terrorists, which called on the president to “use all necessary and appropriate force against those nations, organizations, or persons he determines planned, authorized, committed, or aided the terrorist attacks” on September 11, 2001, in order to prevent “any future acts of international terrorism against the United States.”13 The lesson one might draw from this controversy is that Congress ought to be more careful in drafting resolutions. But better lawyering can go only so far because, as the well-worn phrase “necessity knows no law” reminds us, even the most precisely worded restrictions can legitimately be overridden during emergencies. In such cases lawmakers are typically confronted with the question of whether necessity genuinely warrants a drastic response. And, as Thompson has pointed out, the executive’s control over national security information usually provides the president with a significant advantage in determining whether the circumstances are ripe for a dismissal of previously mandated guidelines.14
So what can Congress do to obtain the information it needs to tailor laws, monitor potential violations, and enforce accountability? It has two powers that it can call upon. It can use its subpoena power to compel officials and agency heads to disclose pertinent information, and it can withhold cooperation by dragging its heels on the confirmation of appointments, the ratification of treaties, and the passage of appropriations until the president complies with requests for information. In practice, though, Congress is usually hard-pressed to employ these powers to the fullest extent. Its subpoena power is hobbled by the problem of nonenforcement because the Justice Department, which is responsible for prosecuting cases involving contempt of Congress, has traditionally refused to press charges against officials who decline to share information with Congress on the orders of the president.15 In theory, Congress could attempt to enforce these subpoenas on its own steam but, as Joseph Bishop has dryly noted, “it has never in the past been willing to push matters to the point of dispatching the Sergeant at Arms to cleave a path through the Secret Service cordon and seize the person of the President, or even one of his subordinates.”16 Congress’s ability to withhold cooperation from the executive is also far from certain. This course of action is vulnerable to swings in public opinion and subject to the usual challenges associated with undertaking collective action.17 It often takes only a little maneuvering by the president to make it painful for Congress to stymie national security initiatives—a point made best by Senator George Aiken’s memorable observation in the heyday of the Cold War: “We all know that when the appropriations bill is pending the Russians in particular become extremely powerful.”18
We have so far been examining the practical obstacles that Congress faces when it tries to compel the president to share information. Suppose, however, that Congress is able to overcome these obstacles and gather the will to challenge the president. Is this the end of the problem? Not quite. Congress faces yet another challenge: since compelling the president to disclose national security information could lead to the release of information that ought to be kept secret in the interests of national security, it would be irresponsible of Congress to utilize compulsion unless it is confident that the president is actually withholding evidence of wrongdoing. But how is Congress to know when this is the case, since it can ascertain whether the withholding was justified only after it sees the information in question?
To appreciate the seriousness of this conundrum, consider William C. Banks and Peter Raven-Hansen’s recommendation that Congress ought to use its power of the purse to obtain some degree of control over national security policy-making. Wishing to encourage congressional oversight, but cognizant of the danger that fiscal restrictions could interfere with the president’s duty to maintain national security, Banks and Raven-Hansen argue that Congress can determine the acceptability of a restrictive appropriation by “weighing the extent to which the restrictions prevent Presidents from accomplishing their constitutionally assigned functions against the need for restriction to promote objectives within the authority of Congress.”19 What Banks and Raven-Hansen do not explain, though, is how Congress might employ this formula when the president alone knows how far such a restriction would hinder the realization of an important national security objective. Unable to verify the president’s claim about the harm likely to be caused by a restrictive appropriation, Congress would be hard-pressed to issue a rebuttal. Congress could try ignoring the president’s claim—a tactic that might help it test the president’s resolve. But to persist in this course of action once the president refuses to submit could have highly undesirable consequences. For example, imagine that the president concludes a nuclear stockpile reduction agreement that favors the United States. Not wishing to embarrass the other side, and to thereby imperil the agreement, the president publicly represents the deal as having achieved a modest outcome that imposes significant burdens on both sides. Unaware of this gambit, a bloc of hawkish senators declares that they will not fund the stockpile reduction program unless they are granted access to the details of the agreement. The president, however, refuses to cooperate because he fears that should the senators leak the details of the agreement—which he thinks they are likely to do in order to prove to their constituents that they are not backing a burdensome agreement—the counterparty would be greatly embarrassed before its constituents and thereby compelled to renege on the agreement. In this case, does it not seem that Congress would be barking up the wrong tree by refusing to fund the stockpile reduction program unless the president shares the information they want?
The conundrum outlined above implies that if Congress wants to compel the disclosure of information without jeopardizing national security, then it ought to use subpoenas or refuse to cooperate only after it has obtained some knowledge of the nature of the information being withheld by the president. But how can Congress obtain such knowledge when the president controls the flow of information? A common response to this predicament is that Congress ought to strike an information-sharing agreement with the president. Consider the proposal that Antonin Scalia, then an assistant attorney general in President Gerald Ford’s administration, floated before Congress in 1975. Scalia proposed that in return for a promise of nondisclosure, the president could informally transfer to Congress the information that he deemed privileged so as to allow Congress to assess the merit of his decision. In the event that Congress determined that the president’s decision lacked merit, it would be obliged to return the documents and to formally seek access to them by using the constitutional powers at its disposal.
So why did Congress not take advantage of this offer? The catch was that under Scalia’s scheme the president could take the view that “even conditional transmittal to Congress might result in irreparable harm.”20 In other words, Scalia’s proposal left the president’s control over the flow of national security information intact. His proposal was not unusual in this regard—by their very nature, information-sharing agreements cannot be enforced by the courts; they rely on comity between the branches. This in turn explains the lack of enthusiasm in Congress for this device. For if the president can stall even a conditional transmittal to Congress, then lawmakers have little reason to expect that a president who has something to hide will hold up his end of an information-sharing agreement. A case in point is the outcome of the so-called Casey Accord of 1986, an informal agreement reached between William Casey, the director of central intelligence, and Robert McFarlane, the national security adviser, on the one side, and Senators Barry Goldwater and Daniel Moynihan of the Senate Intelligence Committee, on the other. Under this agreement, which followed in the wake of uproar in Congress over the previously undisclosed mining of harbors in Nicaragua, Casey and McFarlane promised to keep the intelligence committees in Congress fully informed of covert activities. Yet, as Harold Koh has observed, “within days after accepting these voluntary reporting requirements, the Reagan administration began planning its secret arms sales to Iran.”21
If the analysis presented above is correct, then we must conclude, contrary to Rozell, Devins, Schmitt, and Crockett, that Congress is not well positioned to compel the president to share national security information. This conclusion may seem implausible since Congress has conducted a number of high-profile national security investigations in recent decades, including those by the Church and Pike Committees in 1975, the Iran-Contra Committee in 1986, and the 9/11 Commission in 2004. But, significantly, Congress undertook these investigations only after the press brought potential wrongdoing to public attention, thus allowing lawmakers to feel justified in compelling the president to disclose privileged information.22 In the absence of such assistance, Congress’s powers would arguably have proven far less useful than they otherwise seem. To take just one example, the 9/11 Commission successfully used the threat of subpoenas to obtain access to the President’s Daily Briefing (PDB), which the Bush administration had previously refused to share.23 Yet, as the leaders of the commission have acknowledged, their focus on the PDBs owed substantially to the unauthorized disclosure of the title of the infamous PDB “Bin Laden Determined to Strike in US,” a disclosure that had raised the initial question about the competence of decision-makers in the White House.24
As noted in chapter 1, Congress’s tendency to ride to the rescue on the back of leaks of classified information is troubling for a number of reasons. It has long been recognized that news reports constitute one of the most important means by which Congress keeps track of what the president is doing.25 But news reports based on leaks of classified information are a different matter because such unauthorized disclosures violate laws enacted by Congress itself. To be sure, under certain conditions executive officers may well be justified in disobeying these laws (as we shall see in chapter 5). However, the scope of the cases where such disobedience is justified becomes significantly narrower when secret information is disclosed to the world at large (rather than to overseers in Congress alone, who could address the problem discreetly). As a result, oversight provided by a Congress that relies on leaks will tend to be underinclusive—that is, it will overlook wrongdoing that is not leaked or published out of concern for collateral damage to national security. Furthermore, since leaks typically occur quite some time after wrongdoing has taken place, they often leave Congress with little opportunity to take remedial measures. For example, important disclosures about the politicization of secret intelligence pertaining to Iraq’s weapons program came only after the decision to invade had already been made. And, perhaps most troublingly of all, since unauthorized disclosures are typically made anonymously, Congress may find itself unable to ascertain the motives of the source, leaving it vulnerable to being manipulated by disgruntled or partisan officials. As such, Congress’s dependence on leaks seems to be in some tension with the democratic norm that decisions on vital matters of public concern must be made by individuals who are directly or at least indirectly accountable to the public. Given these concerns, we cannot conclude that Congress is currently well positioned to oversee the employment of state secrecy. Instead, we ought to examine whether Congress can have timely and independent access to the national security information it needs to conduct oversight on its own steam.
The Problem of Executive Privilege
The principal challenge that Congress faces in its endeavor to obtain national security information is that the executive does not believe itself to be obliged to comply with Congress’s requests. As Snider has observed, though Congress routinely passes laws reserving the right to examine intelligence materials as it sees fit, none of these statutes has “been interpreted by the executive to require that all intelligence be turned over to Congress.”26 Is this state of affairs defensible? Should the president be allowed to withhold national security information from Congress?27
So far the subject of executive privilege has primarily been addressed at the level of constitutional doctrine: scholars have focused on whether precedent or original intent or the doctrine of inherent or implied powers entitles the president to withhold information from Congress.28 This scholarship must not be overlooked, because if the president is clearly entitled to control the flow of information to Congress, then we could still ask whether the president ought to have such authority, but there would be, realistically speaking, little hope of wresting this authority away from him. As it turns out, though, the president’s constitutional right to control the flow of information is heavily contested. Moreover, the courts, which usually resolve such interpretive disputes through authoritative rulings, have declined to offer relief in this instance, citing the need, in United States v. AT&T (1977), to “avoid a resolution that might disturb the balance of power between the two branches and inaccurately reflect their true needs.”29
Given the absence of clarity at the constitutional level, let us consider whether there are general reasons in favor of executive privilege. Arguably, there are two justifications to consider here. The first is that the executive privilege provides a means to uphold the independence of the executive branch and to thereby maintain the separation of powers. This claim can be traced back to Attorney General William Rogers, whose testimony before Congress in 1958 drew attention to a well-known passage from Kilbourne v. Thompson (1880) to the effect that separation of powers theory requires “that the persons entrusted with power in any one of the branches shall not be permitted to encroach upon the powers confided to the others.”30 The import of this passage, Rogers argued, was that Congress could not, for instance, pry into judicial deliberations, as this would “be utterly destructive of a free judiciary.” “The same considerations,” he continued, “may be said to operate with respect to an investigation of confidential advice within the executive branch.”31
This argument is not entirely convincing, though. In cases where the information subject to a claim of privilege relates to matters that lie exclusively within the ambit of the executive (e.g., the conduct of a secret military operation), the president’s interest in protecting confidentiality ought to trump Congress’s more general interest in accessing such information. But when the information in question bears on matters that call for Congress to use its constitutional powers—for instance, by funding covert action or authorizing war—then it seems that the president’s interest in protecting confidentiality cannot be conclusive; we must also account for Congress’s interest in having access to the information it needs to ascertain what laws should be passed under the circumstances.32
The second justification for executive privilege is based on the notion that the separation of powers theory vests responsibility for the “stability and security of the nation” in the executive because this branch is best suited to acting with “secrecy and speed.”33 This responsibility in turn, the argument goes, authorizes the president to withhold national security information from Congress whenever this is necessary to protect national security.34 But this argument does not seem entirely convincing either—at least not as it has been presented thus far. In particular, its advocates have not clearly explained why we should believe that Congress is less capable of maintaining due secrecy. Usually the explanation offered cites Hamilton’s claim in The Federalist No. 70 that “decision, activity, secrecy and dispatch will generally characterize the proceedings of one man, in a much more eminent degree, than the proceedings of any greater number; and in proportion as the number is increased, these qualities will be destroyed.”35 To support this observation Congress’s critics often cite the famous letter that Benjamin Franklin and Robert Morris wrote on October 1, 1776, in their capacity as members of the Committee of Secret Correspondence. After informing their fellow committee members of recently arrived intelligence about the willingness of the French to provide military aid, Franklin and Morris argued that it was their “indispensable duty to keep it a secret, even from Congress,” because “we find, by fatal experience, the Congress consists of too many members to keep secrets.”36
Now, to be sure, Hamilton’s theory and Franklin and Morris’s observation are far from trivial. We know that in order to prevent unauthorized disclosures, decisions on matters of state are often made by a few leading officers, with their subordinates having only “compartmentalized” access to information.37 But if the claim at hand is just about the number of decision-makers, then we are hard-pressed to explain why a committee of Congress should not be provided unrestricted access to national security information, because a committee will comprise only a handful of members. This is why when Eric Posner and Adrian Vermeule argue that executive privilege is justified by the concern that “Congress leaks like a sieve,” David Pozen retorts that no one has explained why “the odds of public disclosure increase not linearly but geometrically” when the “circle of secret-keepers is widened to include authorities from another branch.”38
As it happens, there is an explanation as to why the odds of public disclosure increase exponentially when we provide even a few members of Congress with unrestricted access to national security information. The key factor is not the size of the groups that have access to information, but rather how their memberships are composed and structured. In the executive branch, officials owe their position to the president. This makes it less likely that they will have an incentive to reveal information contrary to the decision-maker’s wishes. A committee of Congress, by contrast, invariably comprises adversarial parties, who will usually have a greater incentive to disclose sensitive information in a bid to gain political advantage. The Framer James Wilson makes this point best in his Lectures on Law: “Can secrecy,” he writes, be expected from a body in which “to every enterprise, and to every step in the progress of every enterprise, mutual communication, mutual consultation, and mutual agreement amongst men, perhaps of discordant views, of discordant tempers and of discordant interests, are indispensably necessary?”39
There is a predictable response to Wilson’s charge. It is that Congress can prevent partisanship from having undesirable consequences by strengthening internal safeguards on the handling of classified information and establishing penalties for unauthorized disclosure.40 But this response lacks credibility for at least three reasons. First, lawmakers can find refuge in the Speech and Debate Clause, which shields them from criminal and civil action. This is the provision that Senator Mike Gravel took advantage of in 1971 when he placed the Pentagon Papers on the Congressional Record before the Supreme Court had reached its verdict in that case.41 Second, lawmakers can be subject to disciplinary action only if they are found guilty of having made an unauthorized disclosure. Yet the evidence required to prove guilt is usually hard to come by because lawmakers who make unauthorized disclosures outside the safety of the floor of the House or the Senate tend to do so under cover of anonymity, safe in the knowledge that the president will be hard-pressed to order a penetrating investigation lest he be accused of criminalizing political differences or violating Congress’s dignity and independence. Consider in this respect the events surrounding the House Intelligence Committee’s investigation in 1975 of claims that the CIA and FBI had violated domestic and international laws. The committee produced a final report that the House voted 264 to 124 to withhold from the public. The report was nevertheless disclosed to Daniel Schorr, a reporter for CBS News, who in turn passed it on to the Village Voice, which published it in full. Though the House conducted an inquiry into the source of the disclosure, the culprit was never identified.42 Finally, even when the source is identified, the adversarial nature of congressional politics raises questions about the impartiality of the judges—members of Congress themselves—who are invariably more forgiving of unauthorized disclosures made by members of their own party. Indeed, it is notably the case that congressional resolutions on whether to inquire into unauthorized disclosures invariably proceed along party lines.43 This is what happened in 1992, for example, when Republican lawmakers called for Representative Henry Gonzalez, chairman of the House Committee on Banking, Finance, and Urban Affairs, to be subject to an ethics inquiry after he placed on the Congressional Record classified documents that his committee had obtained during an investigation into the United States’ support for Iraq prior to 1991.44
Contrast the foregoing with the position of officials in the executive branch. These individuals are vulnerable to criminal, civil, and administrative action should they be found responsible for unauthorized disclosures (and, as we shall see in chapters 4–6, such punishments are far from rare). Moreover, as these officials serve at the pleasure of the president, they can be dismissed or transferred or demoted upon the slightest suspicion, making the maintenance of internal discipline that much easier—a point starkly evidenced by President Eisenhower’s well-known announcement at a cabinet meeting that any official who violated the confidentiality he wanted to protect wouldn’t be working for him the following day.45
I have been arguing that executive privilege can be justified on the grounds that the executive branch is better designed to maintain secrecy. At this point it may be objected that, in practice, Congress is actually not prone to making unauthorized disclosures. Heidi Kitrosser has claimed that “Congress is considered to have a reliable track record for nonleakage,” while Pozen has asserted that no one has “marshaled any evidence, even anecdotal evidence,” showing that Congress “leaks like a sieve.”46 Indeed, it may even be objected that when it comes to keeping secrets, Congress is in fact the more reliable institution. For example, Koh has stated that “the executive branch, not Congress, has been responsible for the vast bulk of recent national security leaks,” while Fisher has averred that “there should be little doubt that congressional leaks, as compared to executive leaks, are infrequent and small in number.”47
But claims of this variety are not well founded. As one lawmaker has pointed out, since “only a handful of leaks have ever been traced through investigation to the culpable individual,” the lack of proof of leaks from Congress “hardly establishes that Congress has a good record.”48 Furthermore, claims that the executive branch leaks more than Congress fail to distinguish between two kinds of disclosures: authorized disclosures and unauthorized disclosures. The former are disclosures that are condoned and may even be ordered by the leadership of the branch, because they are intended to “enhance the implementation of a policy.”49 Since control over national security information is presently vested in the executive branch, it might well be responsible for the majority of authorized disclosures. But this fact tells us very little about the relative merits of either of the branches in terms of their ability to keep secrets, since there is no reason to believe that Congress would desist from making equal use of authorized disclosures were it granted unrestricted access to national security information. Hence the argument actually hinges on the question of which branch is better at preventing unauthorized disclosures: disclosures of information that the leadership of the branch has determined ought to be kept secret.
Before we address this question, though, a further distinction is needed. It is no doubt true, as Johnson has noted, that the “most egregious” unauthorized disclosures have come from “officials in the executive branch who sold America’s top secrets to foreign intelligence agents.”50 But such unauthorized disclosures—effectively acts of espionage—are part and parcel of the business of intelligence, a managerial hazard that would exist regardless of who is responsible for managing the intelligence services. Therefore, we should not count this sort of unauthorized disclosure against the executive branch per se. By contrast, the kind of unauthorized disclosure that is relevant to our argument is that which occurs when, to paraphrase Stephen Knott, political actors deliberately reveal information with a view to crippling a policy that they could not defeat through the normal processes of decision making.51 This problem of indiscipline arises from the tendency of individuals to disregard the rules of the institution in the midst of political conflict.52 Since the executive branch manages the national security apparatus, it would not be surprising if the bulk of such unauthorized disclosures could be traced back to its officers. But if we are to fairly compare the ability of the branches to prevent unauthorized disclosures, then we need to know how Congress would perform if its access to national security information were unrestricted.
Now, for obvious reasons, we are unlikely to be able to resolve this question empirically. However, two observations are worth making. First, given that the Continental Congress did have wide access to national security information, we should not overlook the Framers’ dim view of that body’s capacity to maintain discretion. Furthermore, though the modern Congress has limited access to national security information, its inability to punish even blatant violations of its own rules suggests that granting it wider access to national security information could cause grave harm. Consider, for example, the actions of Senator Burton Wheeler, a staunch isolationist opposed to the United States’ entry into World War II, who in 1941 revealed the United States’ plan to ship troops to Iceland.53 Wheeler defended his decision to reveal news of troop movements, made at a time when German submarines were aggressively patrolling the North Atlantic, as an exercise of his “right to free speech.”54 Another widely cited example comes from 1974, when Representative Michael Harrington disclosed to the press secret testimony given to an intelligence subcommittee by William Colby, the director of central intelligence, about the United States’ efforts to overthrow Salvador Allende of Chile. Though Harrington had initially agreed to keep Colby’s testimony confidential, he later claimed that “he felt he had a greater duty to release the information.”55 Consider as well the case of Representative Clement Zablocki, the chairman of the House Foreign Affairs Committee, who in 1981 leaked to Newsweek information about a planned covert operation against Libya.56 Zablocki suffered no disciplinary action because, in the words of Edward Boland, then chairman of the House intelligence committee, “leaks were epidemic.”57 Then there is the case of Representative Les Aspin, the chairman of the House Armed Services Committee, who in 1987 disclosed to the press information gleaned from a briefing on planned military maneuvers in the Persian Gulf.58 Aspin defended this disclosure with the argument that he had not been informed that the information was classified.59
These examples of reckless behavior by lawmakers in the face of disagreement with the president, and with their colleagues in Congress, shed light on why observers have argued that releasing information to Congress is “tantamount to setting a ship afloat at sea without benefit of motor, rudder, or sails.”60 To be sure, the executive branch is by no means immune to the problem of indiscipline. On the contrary, as we shall see in chapters 4–6, the executive branch faces disciplinary challenges of its own. But the president’s power over his subordinates—and his willingness to use it—allow us to believe that the executive branch can punish (or at least seriously try to punish) those responsible for making unauthorized disclosures. By contrast, Congress’s structure and composition give little reason to believe that its leadership could do likewise. Indeed, what is most striking about the examples cited above is how often lawmakers have openly defied Congress’s norms and regulations concerning secrecy, a practice suggesting that they do not greatly fear the penalties associated with indiscipline (indeed, none of the individuals mentioned in these examples suffered any meaningful sanctions). Contrast this with the preference that disgruntled officials in the executive branch have for anonymity, and the heavy hand of the president comes into sharp relief. To put it differently, even though both branches suffer unauthorized disclosures, one seems more willing and able to patch the holes. Herein lies the justification for the executive privilege.
Should We Abolish Executive Privilege?
We have now established that the executive’s greater capacity for discretion warrants granting the president the right to withhold national security information from Congress. This conclusion may give rise to the objection that we ought to allow Congress unrestricted access to national security information even if lawmakers are more prone to making unauthorized disclosures. This is because under certain circumstances—for example, when the country is faced with a decision to wage war—the benefit of allowing lawmakers unrestricted access could outweigh the harm that might be caused if they disclosed such information. If we fail to provide Congress with unrestricted access in such cases, the objection goes, we will be guilty of treating national security as a “trump” rather than as only one of a number of important interests that need to be taken into account, including, critically, the Framers’ desire to prevent the concentration of power.
This objection is ill-conceived. The justification offered above on behalf of executive privilege does not preclude a balancing of the public’s interest in national security on one hand and oversight on the other. It implies only that the responsibility for balancing these interests ought to be vested in the hands of the president. To see the logic behind this arrangement, consider the following: while we can agree that the benefit of allowing lawmakers unrestricted access to national security information could outweigh the risk that they may disclose this information at great cost to the nation, how are we to know when this is the case? Obviously, this question cannot be answered without first ascertaining the value of the information in question. Needless to say, we cannot ask Congress to judge the value of this information, since this would prematurely expose the information to the higher risk of disclosure associated with congressional access, thereby defeating our attempt to balance interests. What we need, then, is an intermediary. And this is precisely what an executive privilege accomplishes: it allows the president to withhold information in instances where he judges that the public’s interest in preventing unauthorized disclosure outweighs its interest in allowing Congress access to that information. A good example is the so-called Canadian Caper, a CIA covert operation that led to the rescue of American hostages hiding in the Canadian embassy in Tehran in 1979. In this case, President Jimmy Carter notified Congress about the covert operation only after the three-month-long operation was complete—even though the law required more timely notification—because the Canadian government premised its cooperation on Congress’s being kept out of the loop, as it feared that a brash or unguarded lawmaker might expose the operation.61 In this instance, the existence of a discreet intermediary—namely, the president—made it possible to strike a measured balance between secrecy and oversight.
Nonetheless, we cannot ignore the fact that permitting the president to serve as an intermediary comes with a substantial risk. The president could use his control over the flow of secret information to disarm opposition from Congress by sharing with it only those pieces of secret intelligence that support his preferred policy choices.62 In such cases, Congress may well exercise its powers by, for example, issuing declarations of war or appropriating funds, but its actions will not be informed—its situation will be akin to that of a president granted the power to veto a bill, but not to examine its contents. Under the circumstances, we would be justified in curtailing executive privilege should we determine that this privilege is likely to engender abuses of power that will ultimately prove more harmful than the unauthorized disclosures it is meant to prevent.
Although it is hard to prove that the balance of risk actually works out this way, let us assume for the sake of argument that allowing the president to withhold information from Congress is in fact more dangerous than allowing Congress to have unrestricted access to it. Even so, would curtailing executive privilege greatly reduce the chance that secrecy may be employed to conceal wrongdoing? I have been arguing that lawmakers cannot be counted upon to oversee the president so long as the executive controls the flow of information. Presumably, then, once Congress is provided unrestricted access to national security information, the problem of ascertaining whether secrecy has been employed responsibly should be eliminated. Upon closer inspection, though, this turns out not to be the case.
To see why providing Congress with independent access to national security information will not eliminate fears about the misuse of secrecy, we must account for the steps that Congress would need to take in order to merit such access. Given its structural disposition toward indiscipline, it would arguably be irresponsible for Congress to seek independent access without first attempting to reduce the risk of unauthorized disclosure. How might this be achieved? The recommendation offered by scholars such as Koh and Kitrosser is that Congress should delegate the oversight of national security matters to a “core group of members” comprising a handful of its highest-ranked officials.63 Given our previous discussion, we know that this recommendation misses the point; it fails to recognize that the risk of unauthorized disclosure is less a function of the number of people who have access to a given secret and more a function of the degree of partisan conflict. Needless to say, there is not much we can do to address the issue of partisanship, since we cannot eliminate members who belong to political parties opposed to the president without also undermining the credibility of the oversight provided by this core group. Therefore, our hope must be that members will be deterred from making unauthorized disclosures by the threat of severe legal and political sanctions. The credibility of this deterrent is rendered somewhat doubtful, though, by the fact that unauthorized disclosures can be made anonymously. Indeed, there is the possibility that members of the core group who belong to the president’s party could make anonymous disclosures in order to discredit lawmakers from the rival party. This possibility is less far-fetched than it sounds. For instance, in 2002 an unauthorized disclosure by a member of Congress’s joint inquiry into the events leading up to the 9/11 terrorist attacks prompted the Bush administration to threaten to cease to cooperate. A prolonged investigation by the FBI eventually revealed the source of the disclosure to be a Republican, Senator Richard Shelby.64 This turn of events led Senator Bob Graham, the cochair of the joint inquiry, to publicly wonder whether the unauthorized disclosure had been intended to sabotage the joint inquiry (presumably by giving the Bush administration a pretext to end its cooperation with an unwelcome investigation).65
Let us nonetheless assume, for the sake of argument, that restricting the membership of this core group will increase the probability that “cheaters” will be caught to a level high enough to bring unauthorized disclosures under control. The question to ponder, then, is this—how effective is such a core group likely to be at preventing the misuse of secrecy? Presumably, the group’s members will now have the information needed to determine whether Congress ought to support or oppose the president’s policies. But can we be confident that they will act dutifully when their own conduct as overseers is shielded from public view?
One problem is that the secrecy accompanying the activities of this core group may leave its members unable to explain to the public why they wish to block or investigate the president’s policies or decisions. Hobbled in their ability to rally public opposition to policies that they view as harmful or unwise, the group members may be rendered little more than mute spectators—as Senator Jay Rockefeller discovered in 2003 after being briefed, in his capacity as vice chairman of the Senate Select Committee on Intelligence, about the NSA’s warrantless wiretapping program. According to Senator Rockefeller, because he was not allowed to discuss the program with his staff, outside experts, or colleagues, the only way he could voice his concern about the propriety of the program was to complain privately to Vice President Cheney.66 This sort of predicament is far from unprecedented—indeed similar complaints can be traced back to the 1970s, when intelligence oversight procedures were first put in place.67
The prospect that the core group may be bulldozed by the president may lead to the demand that it be allowed to overrule the president and to share its views with Congress or even the public.68 However, before we accede to this demand, we ought to ask what will prevent the group from misusing such a prerogative? This question will of course be posed most acutely when a majority of the members of the core group come from the same political party as the president, an arrangement that naturally raises fears of collusion. These fears do not seem misplaced when we consider, for instance, the criticism of the Republican-dominated intelligence committees for their willingness to turn a blind eye to the Bush administration’s more controversial national security initiatives.69
Some scholars have responded to the prospect of ineffectual oversight by calling for oversight committees to be chaired by members of the rival party. Bruce Ackerman, for instance, recommends that during an emergency the president ought to be allowed to exercise extraordinary powers subject to close and continuing oversight by Congress. Noting that Congress cannot provide the requisite degree of oversight “if it is at the mercy of the Executive for information,” Ackerman instructs that during an emergency “members of opposition political parties should be guaranteed a majority of seats on oversight committees,” and that the president should be required “to provide the committees with complete and immediate access to all documents.”70 This arrangement “puts the government on notice,” Ackerman writes, “that it cannot keep secrets from key members of the opposition and serves, without more, as an important check on the abuse of power.”71 This proposal has been backed even more strongly by Kim Scheppele, who has argued that adversarial oversight ought to become a permanent feature of the American institutional landscape (rather than being confined merely to emergency settings). In particular, she has urged the United States to consider following the example set by Germany, where the system of oversight is premised on the idea “that the opposition parties must be able to check that majority parties are not using the intelligence services for their own political purposes.”72 The German political system creates such a “political check,” Scheppele reports, by requiring the government to “report all intelligence activities” to the Parliamentary Control Commission, which is chaired by the majority and minority parties on a rotating basis.73
Upon closer inspection, though, these proposals turn out to be problematic in a number of respects. First, both Ackerman and Scheppele assume that lawmakers will have the expertise and political capital available to enter into bruising contests over intelligence matters with the executive, especially during times of crisis. However, as Martha Minnow has pointed out, the evidence suggests that congressional deference to the executive branch actually increases during times of crisis.74 Moreover, even if oversight committees are stocked with experienced and unflappable lawmakers, can officials be trusted to divulge incriminating information to lawmakers, especially when they are appointed by, and likely owe loyalty to, the president?75 It is worth recalling how often Congress has been misled by executive officers. In 1977, for example, Congress discovered that Richard Helms, then the director of central intelligence, had misled it during his testimony over the CIA’s activities in Chile.76 In 1986 Congress discovered that Colonel Oliver North had misled it with regard to the Iran-Contra affair.77 Congress’s testy relationship with William Casey, the director of the CIA during the Reagan administration, is legendary. As Representative Mineta once said in exasperation, “[I]f you were talking to Casey, and your coat caught fire, he wouldn’t tell you unless you asked about it.”78 This pattern does not seem to have faded with the passage of time. Over the past decade both George Tenet and General Michael Hayden have been accused of misleading Congress during their tenures as directors of the CIA and the NSA, respectively.79 Nor does the problem of noncooperation seem unique to the United States. In 2011–12 German investigators discovered that members of the intelligence services had hidden documents from overseers and utilized espionage and surveillance techniques beyond those approved by the Federal Constitutional Court.80
These practical concerns are not, however, the primary reason to be doubtful about Ackerman’s and Scheppele’s proposals, because these issues could conceivably be addressed by Congress’s vigilance in approving appointments to senior positions in the intelligence community. Let us focus therefore on a deeper problem: namely, that concerns about the quality of oversight do not fade away even when a majority of the members of the core group come from the party opposed to the president. This is because the majority could exploit its position for partisan purposes—for example, by selectively revealing national security information that furthers its own agenda or by taking one position in a closed session and then another in public when the political winds turn unfavorable.81 In this setting, the shoe will simply be on the other foot, as it will now be the president who is left unable to fully explain why the released information should be discounted. An illustration of this problem is provided by controversy surrounding the findings of the Rumsfeld Commission, which was created in 1998 after Republicans in Congress, who wished to see the United States pursue national missile defense, accused the Clinton administration of forcing the CIA to downplay the ballistic missile threat posed by rogue states like Iran and North Korea.82 The Rumsfeld Commission, which demanded and was subsequently given unrestricted access to classified information, concluded that the available intelligence analysis was faulty, and that the United States faced a far greater threat than the Clinton administration was willing to admit.83 The Clinton administration responded by challenging this “alarmist” interpretation of the available intelligence. But since external observers could not verify the claims made by either side, the entire dispute quickly turned into little more than a public relations battle. And though the Clinton administration quickly lost this battle, the passage of time has shown that its measured claims were well founded.84
The quality of oversight provided by a core group is also brought into question by the fact that the secrecy surrounding their operations makes it unlikely that they will be held accountable for regulatory failures. Consider, for example, the aforementioned case involving Senator Rockefeller. Did the secrecy demanded by the Gang of Eight (the leaders of the parties in both chambers and the ranking members of the intelligence committees) provide Senator Rockefeller with the opportunity to turn a blind eye to a policy that would have deeply upset his constituents, or was he, as he later claimed, an unhappy but silent observer of the Bush administration’s policy? What we do know is that at the time, Senator Rockefeller placed a copy of his letter to the vice president in his personal safe as a testament to his unhappiness with the Bush administration’s policy. But a skeptic might argue that this move merely ensured that Senator Rockefeller would have political cover no matter which way the political winds blew: had the warrantless wiretapping program produced a stunning success in the war on terror, he would have avoided the awkwardness of having publicly opposed the program; conversely, if it proved deeply controversial (as it eventually did), then he would have a get-out-of-jail-free card at hand. The point of this example is not to impugn Senator Rockefeller’s motives. Rather, its purpose is to show that when we rely on a core group like the Gang of Eight to oversee national security matters in secret, we are invariably left with more questions than answers about the conduct of its members.
Contrary to what Ackerman and Scheppele assume, then, the use of secretive oversight committees is unlikely to greatly improve our confidence in the quality of oversight on offer. There is a deeper point here that is not always recognized in discussions on the subject of oversight: to the extent that someone must decide what information is to be shared or concealed, it is always possible that such a person or committee will abuse this authority. It is often assumed that a committee will be less vulnerable to committing such abuse. But why should this be so? A committee will have to make decisions via a decision-making rule such as majority or super-majority voting. What is to prevent a factional or corporate interest from emerging dominant within the committee? It is not possible to cancel out the effect of factions by expanding the size of the committee. Nor should we expect or encourage individuals or minorities to violate the rules of the committee and make ad hoc unilateral disclosures to Congress or the public—were this to happen, the whole purpose of entrusting the decision to a group would be defeated.
It appears, then, that taking the final say over information sharing away from the president and vesting it in a congressional committee serves only to re-create the opportunity for the abuse of state secrecy. This realization may well prompt proponents of checks and balances to argue that the problem of destructive partisanship can be countered by entrusting the final say on information sharing to a court or an independent panel composed of “nonpartisan experts,” who would decide when the president should be allowed to withhold national security information from Congress. For example, John Orman has proposed that Congress should legislate categories of information that the president can legitimately withhold from it, and that in the event of conflict over what is to count as a legitimate secret, “the courts should act as the final arbitrators over all uncategorized information as well as all disputes about the proper categorization of information.”85 Similarly, William Weaver and Robert Pallitto have called for the judicial arbitration of executive privilege claims on the grounds that the “courts must help establish what the separation of powers is, rather than using the phrase itself as a justification for staying out of the fray altogether.”86
This line of thought finds some support in United States v. AT&T (1977), a case involving a dispute between Congress and the Department of Justice over access to classified documents concerning the FBI’s use of warrantless wiretaps. Although the Court of Appeals for the District of Columbia declined to address the constitutional aspect of this “portentous clash” between “the executive branch asserting its authority to maintain tight control over information related to our national security, and the legislative branch asserting its authority to gather information,” it did volunteer to fashion a compromise on the grounds that judicial abstention would not encourage an “orderly resolution of the dispute.”87 The compromise it forged was to allow select congressional staffers to review a random sample of the relevant documents on the condition that the FBI could, subject to an in camera review by the court, substitute the randomly chosen documents with others, should any of the original choices be too sensitive to expose to the higher risk of disclosure associated with congressional involvement.88 Although the proceedings in AT&T did not exactly vindicate the utility of judicial involvement—it eventually took three cases, two courts, and more than eighteen months before the two sides could be coaxed into reaching a contentious settlement—it is still worth considering whether this precedent ought to become more the norm.
The first question we ought to ask here is what reason is there to believe that judges or independent experts are likely to prove reliable arbitrators? The answer, presumably, is that we expect these actors to make decisions based on reasons rather than interests. But what are the relevant sorts of reasons in the present context? Arguably, disputes about information sharing turn on the question of which branch has the most reasonable claim to the information under the circumstances. That is to say, the appropriate answer in a particular case would presumably change if the court believed that a congressional committee seeking highly sensitive national security information contained an alcoholic prone to indiscretions, or if it believed that a president refusing to share information was actually shielding evidence of criminal actions.89 If so, then, as Schmitt has observed, to require the courts to act as arbitrators is to ask them to exercise “political wisdom par excellence,” even though they are usually not in a position to “assess adequately the political conditions and nuances which may infect a dispute.”90 This point shines through in AT&T where the D.C. Circuit Court initially asserted that witnessing negotiations between the two parties had helped it ascertain the “relative magnitude” of the relevant considerations—Congress’s need for information, the likelihood of an unauthorized disclosure, and the magnitude of resultant harm to national security—but then later concluded that briefs on the sensitivity of the material in question may still be of “significant help.”91
Furthermore, even if the courts or an independent panel of experts are willing and able to take on the role of arbitrators, the notion that decisions concerning the distribution of national security information can be made apolitically is itself highly problematic.92 The difficulty lies in the nature and the import of the decisions this court or panel will have to make. Consider, for example, a case where an independent panel of experts has to decide whether to release information pertaining to nuclear proliferation by another country. In making its decision, the panel will want to take into account the potential costs (e.g., the harm that would be caused by the revelation of sources and methods of intelligence gathering) as well as the potential benefits (most notably, an informed Congress). It is hard to imagine that such costs and benefits can be estimated or indeed compared objectively. On the contrary, a great deal will depend on the political beliefs of the adjudicators. Now if the matter at stake were politically unimportant, the beliefs of these adjudicators would also be relatively unimportant. But when the decision in question significantly affects the ability of the political branches to justify controversial policies and to become aware of potential violations of law, the beliefs of these adjudicators will matter enormously. Consequently, it will surely not be long before political parties begin to take interest in the sorts of individuals appointed to such positions, the end result being a less-than-disinterested bench or panel.
This is not all. If a court or panel is to prevent unauthorized disclosures as effectively as the executive branch does, then it will likely have to create special forums where one or very few judges or experts will examine the relevant materials in camera and ex parte. But if we take away the opportunity for external observers to study the basis of the decision reached by these judges or experts, how can we ascertain their disinterestedness? Put another way, what reason will we have to believe that a panel of judges or experts acting in secret will exercise their discretionary powers any differently from their counterparts in the political branches?93 For instance, what are the chances that an “independent” panel would have disagreed with President Bush’s decision to keep information about the warrantless surveillance program from Congress if Presidents Reagan or George H. W. Bush had previously nominated to such a panel individuals who shared the worldview of Chief Justice William Rehnquist or Justice Antonin Scalia? In the end, then, one might expect to see Congress granted or denied access to national security information depending on shifts in the ideological composition of this court or panel.
The arguments made above bring us back to the conclusion reached earlier concerning the advisability of creating a core group. That is, it makes sense to disturb the president’s control over national security information only if the proposed alternative—be it a committee, a court, or a panel—is more certain to serve the public interest. This is not to argue that entrusting the final say over information sharing to a body other than the executive branch is certain to have no positive effect whatsoever. Such a body may well have a less immediate incentive to misuse its control over national security information, since it cannot hope to use this authority to “unlock” additional powers—for example, by manipulating Congress into declaring war. But the problem is that we will not know if and when this is the case, that is, whether the members of this bench or panel have been able to resist being swayed by their partisan affiliations. This is especially troubling given that the political significance of the subject—the distribution of national security information—makes it difficult to imagine that there could even be such a thing as a disinterested intermediary. In other words, under conditions of state secrecy, we will invariably lack a good and sufficient reason to trust any “secrecy regulator,” since we will not have access to the information necessary for rational trust.94 This may be contrasted with the case of a regulator of an industry, where the availability of information allows the regulator’s disinterestedness to be critically evaluated rather than assumed.
What Is the Alternative?
I began by making the case that Congress is not well positioned to oversee the employment of state secrecy because even when it has the willpower to challenge the president, it cannot easily obtain the information it needs to conduct oversight. This does not mean that the current system of congressional oversight is pointless. As Minnow has rightly argued, even limited briefings to lawmakers can serve as a “black box” that can keep a “record of all that is done” so that there are “subsequent occasions for evaluation and review.”95 For this reason it makes sense to take the small but important steps that will incentivize overseers to take their roles more seriously. For instance, we can agree with Kitrosser that requiring members of the Gang of Eight to offer written responses to secret briefings will lessen their ability to pass the buck at a later date and may therefore incentivize them to push officials for explanations at the appropriate time.96 Similarly, we can agree with Kathleen Clark that Congress ought to ensure that lawmakers are able to have their lawyers and staff members present at closed-door briefings; otherwise lawmakers will be hard-pressed to evaluate what information is shared with them.97
Nonetheless, we ought to be realistic about what congressional oversight can achieve even if these small steps are taken. Given the president’s stranglehold over the flow of national security information, there is little reason to believe that lawmakers will be able to take the lead in uncovering policies and actions that the president has decided to conceal. It is now clear why it would be unhelpful to respond to this obstacle by calling for the abolition of executive privilege. Such a step would demand a sacrifice of either secrecy (if the authority to decide what information is shared with Congress is transferred to a committee composed of adversarial parties) or expertise (if this authority is transferred to an ad hoc bench). Furthermore, because it would concentrate regulatory authority in a different but still inscrutable institution—such as a secretive bench or congressional committee—such a step would beg the question of how we can ascertain whether the new regulators of state secrecy are in fact behaving responsibly. Thus abolishing executive privilege would likely hurt national security without really bolstering regulatory credibility.
It is important to recognize that leaving executive privilege intact does not render Congress helpless as an overseer. Though the president can utilize his control over the flow of secret information to thwart Congress’s “police patrols,” we have seen that he cannot easily prevent his subordinates from sounding “fire alarms” (to use the terminology coined by Mathew McCubbins and Thomas Schwartz).98 Therefore, Congress will often be able to rein in the presidency. The question is—at what cost? To answer this question, we must examine whether we can regulate unauthorized disclosures: the more we can minimize the risk that such disclosures will violate key democratic norms, the less uncomfortable we will have to feel about Congress’s dependence upon them.