"Eyes on Spies," by Amy B. Zegart

Published by Hoover Institution Press (2011)

Reviewed by Kenneth Anderson

Amy B. Zegart’s splendid new book, Eyes on Spies, is a brisk and brief discussion of why Congressional intelligence oversight is so persistently inadequate. The book, it bears noting at the outset, is political science and not law. It hardly mentions the enabling statute of the CIA, USC Title 50, for example. It barely mentions and nowhere gives even a potted description of the way in which Congressional oversight operates now under the law. That is not its purpose. It offers, rather, is a serious structural and institutional account for why Congress has so few incentives to take up intelligence oversight in any strong way, backed up with ample evidence that indeed this is the case. The book’s very considerable value is that the day-to-day weaknesses of intelligence oversight cannot be understood without understanding the structure of institutions driving those weaknesses. It is an important read for anyone concerned with national security oversight.

Accountability of the executive and its agencies in national security affairs has many flavors but, at its core, it all ultimately comes down to two: the judiciary and the Congress. When it comes to the most contentious activities of the national security executive--detention, interrogation, military commissions, surveillance, targeted killing of American citizens, to take the most prominent controversies of the past decade--debate divides between two camps. There are those who insist that only judicial review comports with the Constitutional rule of law. And there are those who insist upon Constitutional limits to the judicial role in the Republic and who claim that in important national security matters, particularly abroad, the buck stops with the political branches and oversight rests with the Congress.

This basic “who” question of oversight for the executive runs to fundamentally differing visions of accountability. It cannot be resolved except by court decisions that, by and large, are not seen as establishing institutional settlement; the normative frameworks are too far apart for that. And the terms of debate here are essentially normative. Those embracing judicial supervision and review of these various national security mechanisms--if not their outright rejection by the courts--argue that the rule of law insists upon review by a non-political actor. Those calling for congressional, not judicial, oversight have equally normative views, drawn from democratic conceptions of a self-governing community, and the roles of the political branches in securing the safety of the commonweal. The argument is intrinsically normative on each side, even if the sides sometimes pause to talk about concerns over effectiveness, resources, expertise, and institutional capacity.

It is also an argument over authority in a constitutional republic. Accountability and authority are the twin pillars of democratic governance, and although at the most abstract level they reinforce each other, at the practical political level they are in tension. An increase in one is a decrease in the other. Authority is required for leaders decisively and coherently to act; accountability is required to ensure that the actions they take are legitimate and therefore exercises of authority and not merely of power alone. But democratic accountability in a practical sense is inevitably a form of second-guessing, undermining authority in the present moment, even if it reinforces it over the long run.

Enter Zegart to this argument. For her, the normative argument is largely beside the point, and the ideals of authority reinforced by accountability rendered effectively moot by the brutal facts of the structural incentives governing Congress. A leading political scientist with a distinguished track record in intelligence studies, Zegart is a national security scholar at Stanford and a senior fellow of the Hoover Institution (which published this volume as part of its series of “short” single-topic books, modestly priced and readable on a plane flight). Students seeking an outline of the law, formal and informal legal mechanisms of intelligence oversight, will have to go elsewhere. Current issues running to the substance of what is taken up by Congress are not taken up at all. What this book does convey is discouraging news for those wanting to see Congress as the natural seat of accountability in intelligence. From the vantage point of social science, Zegart explains how Congress does a poor job of oversight of intelligence issues and lays out the many daunting institutional reasons why this is so.

Her methods are rational choice and public choice theory, on the one hand, and impressive empirical studies of how Congress actually behaves, on the other—studies that are sufficient to show what the theory of incentives would predict. Zegart starts with a review of the literature on congressional oversight in general; she discusses two basic theories of oversight, the “police patrol” and the “fire alarm.” The policing model says that for certain functions, Congress is constantly patrolling executive activities, whereas for others, Congress outsources informally to other actors (including lobbyists) and then intervenes when the fire alarm occasionally goes off. Either way, a substantial part of the oversight literature suggests that congressional oversight, either in the way it designs agencies or undertakes oversight, hits the “Goldilocks” mean of not too hot, not too cold, but just right.

Whether that sanguine conclusion about the efficiency of congressional oversight is generally true or not, Zegart says flatly that it is not true of intelligence oversight by Congress. On the contrary, at least as measured by congressional activity--such as holding hearings or allocating staff resources--intelligence oversight is paltry by comparison to other areas of oversight, such as banking and finance, armed services, and many others. Indeed, Zegart’s research shows that intelligence oversight is nearly at the bottom of the heap in terms of resources and attention, hearings and bills, by comparison to almost any other activity. (Ethics does still worse.) This is as true of the past ten years as it has been of the preceding decades. Despite the undeniable importance and growing reach of intelligence activities, and the steep growth of the intelligence budget and community, intelligence oversight remains practically as moribund as ever.

One might express skepticism of the standard political science measures of oversight effectiveness used here. After all, holding hearings and introducing or passing bills will be be a feature of “good” oversight, but it will also be a feature of “bad” oversight. Activity is not by itself a guarantee of quality. But quality oversight requires a certain minimum level of attention, and comparison to other congressional oversight interests gives an important indicator of institutional attention. And Zegart offers a simple explanation for the inattention—one drawn from public choice theory. The end-goal for members of Congress is re-election, and intelligence is a national good, not one that contributes in any obvious way to re-election in a local district. If anything, service on the intelligence committees is a distraction from issues that matter directly to voters at home; it is far more secret than other activities and much of it cannot be discussed, even as a point of prestige. And while Congress has evolved mechanisms by which committee work necessary for the internal functioning of the institution can be rewarded despite no obvious and direct connection to a member’s district, those prestige and leadership-rewarded mechanisms do not work well in the case of intelligence.

This scholarship does a fine job of bringing together a plausible deductive theory about political incentives--together with data showing persuasively that the incentives account is true and that whatever validity the Goldilocks model of oversight might have for many executive activities, it does not work for intelligence. Beyond this, Zegart goes on to describe additional, intelligence-specific problems. Earlier reforms of the process took their toll, principally the 1970s-era term limits on intelligence committee service. Although finally abolished for the Senate, these persist in the House of Representatives. While they serve the 1970s-era reformers’ goal of insuring that members do not get too cozy with those over whom they exercise oversight, they also have the effect of insuring that committee members do not have time to develop true expertise in the way members do on other committees. Combined with the limited resources devoted to staff, and the secrecy requirements that apply even where staff possess the requisite security clearances, the result is an oversight process that operates at a special information disadvantage in relation to the executive agencies.

Zegart isolates one other major failing in the existing oversight process. This is the fragmentation of budgetary authority as between the intelligence authorizing committees and the appropriating committees. Whatever sense this fragmentation makes for other public business, it makes little sense for the intelligence budgetary process. Despite efforts by Senator John McCain and a few others to unify the system, efforts to oversee an increasingly large budget that includes such things as secret satellite programs running into the billions are easily gamed by agencies or private parties. Budgetary fragmentation ensures that executive agencies and private parties have an overwhelming informational advantage, as well as the availability of a budgetary end-run to the appropriators.

None of this is positive news, though Zegart makes a valiant effort to extract a positive reform agenda out of it all. Intelligence oversight will never be done well, she says, but it can be done better. Her final structural reform take-aways are to remove lingering House term limits; consolidate budgetary authority; and strengthen staff support. The problem is that the baseline incentives from the standpoint of individual members of Congress and their re-election imperatives will remain fundamentally unchanged. Indeed, the incentive structure is even worse than these reforms address, as she notes. At bottom, oversight over intelligence activities means very little upside credit for the demanding, yet secret, task of monitoring activities conducted in secret. Where there is upside credit to be taken, information is leaked by the administration, which takes the gains for having acted. It is hard to get credit for having asked probing questions of an activity one is not free to disclose to the press or public--even if leaked by the administration. It is hard to get credit for having asked good questions in secret sessions and, in the end, not objected.

The downside blame, on the other hand, sticks to committee members. Everyone remembers Nancy Pelosi’s twisting and squirming to avoid admitting she had been briefed on waterboarding. A member of Congress, particularly one among the senior leadership of either party, has to see nearly all risk and very little reward in taking part in classified briefings on potentially controversial things that the executive might determine it must carry out. It is frankly better not to know, and even better not to serve on the committee at all. Although Zegart’s book does not take up this issue, anecdotal evidence suggests that members of committees prefer to be briefed relatively little, without a any recorded or videotaped record of the briefings, and to be briefed as much as possible in generalities. None of this is surprising, given the incentives involved. It does not even yield earmarks, as Zegart’s quite striking data shows.

If there is any slender reason for hope that intelligence oversight might take firmer hold, it lies in the incentives not of the Congress, but of the executive itself. The Bush administration never recognized this, in its obsession with executive authority exercised alone, but the biggest reason for pressing congressional oversight is precisely in order to share the downside risks. If something goes wrong, and it will--or if the President decides that he must take some extremely controversial action, then the best the administration can hope for is to show that to the extent possible, the activity was discussed in concrete terms, in advance, with Congress. Hang together or hang separately. There seems to be no doubt that Leon Panetta, with his long experience of Congress, understood this implicitly when he became CIA director; some press and insider accounts suggest that he insisted that Congress would not only be offered the required briefings, those briefings would be much fuller and designed to ensure that there could be no deniability on the part of senior Congressional leadership as to what controversial actions the CIA had taken.

This issue goes beyond Zegart’s discussion, to be sure, and were she to undertake a study of executive behavior here, the results might well be an equally daunting set of executive branch disincentives to sharing responsibility for difficult, secret, controversial intelligence policies between the political branches. But if there is hope for more effective congressional intelligence oversight, much of the motivation will lie with an executive concerned to show the country that the two political branches are on the same page. This is more than simple institutional self-interest and strategic behavior. Consider the oversight issues as the United States moves to a counterterrorism policy built around discrete and targeted uses of force against particular and identified individuals, some of whom will be Americans. The enemy will become increasingly individualized and targeted through new technologies of surveillance and identification, such as facial recognition technology and miniaturized surveillance robotics. As the US draws down troops from Afghanistan, the civilian CIA might well be left behind in charge of shadowy local forces to gather intelligence and conduct operations to prevent transnational terrorists from re-taking root, and to serve as a base for covert attacks upon terrorists in an increasingly hostile Pakistan. So-called Title 10 operations of the military are already merging with Title 50 intelligence operations, a merger of military and civilian intelligence operations; it makes a lot of functional sense, but also raises new questions about the nature of Congressional oversight and lines of political accountability.

Zegart’s account of the institutional and structural barriers to effective oversight as things stand now, let alone as they evolve along these lines, is timely and sobering. Uses of force against transnational terrorists and non-state actors over time might well become “intelligence-driven uses of force”--colloquially, an increased reliance upon “covert action,” small-scale uses of force in the shadows of states and largely unnoticed and unreported to the public, except when an operation blows up spectacularly. How should accountability here be structured? How can the political branches set out a path of prudent use of these new technologies and strategic capabilities for uses of force that are something less than overt conventional war?

One could make a social science argument for the inadequacy of the judiciary to perform these functions, and no doubt someone will do so one of these days. Either way, no one should find satisfaction in knowing that there might well be no branch of government structurally incentivized to perform well the vital role of intelligence and secret national security oversight. A wise executive branch looks beyond its short-term freedom of maneuver and sees that over the long term, it must have authority and not merely power. Authority in a democracy is power invested with legitimacy, and that legitimacy derives from accountability; this is so whether or not we believe that accountability must be provided in any given matter by judicial or legislative oversight. Robust mechanisms of accountability are what finally and, for the long run, preserve the legitimacy and authority of the executive. That kind of wisdom, however, is far from a structural feature of our political system.

Kenneth Anderson is the Lawfare Reviews Editor. (Full disclosure: Like Jack Goldsmith and Benjamin Wittes on this blog, Anderson is a member of the Hoover Institution’s Task Force on National Security and Law, which provided financial support for book publication; he commented on an early workshop paper by Zegart on this topic several years ago.)