Separation of Powers Response 126 Harv. L. Rev. F. 188

A Reply to Professor Katyal

Responding to Neal Kumar Katyal, Stochastic Constraint, 126 Harv. L. Rev. 990 (2013)

Response To:


Power and Constraint is a response to the conventional wisdom, widespread in the legal academy, that in the national security context we live “in an age after the separation of powers, and the legally con-strained executive is now a historical curiosity.”1 The book argues, to the contrary, that the executive branch under two post-9/11 terror presidencies has been deeply constrained by law (and politics). It shows how Congress and the courts — traditionally weak checks on wartime presidents — pushed back hard against Presidents Bush and Obama and imposed unprecedented restrictions. It further shows that these traditional forces were supplemented by a vigorous press aided by advances in digital technology; by lawyers, inspectors general, and other “watchers” deployed in novel ways inside the presidency; and by consequential nongovernmental organizations (NGOs) that brought successful campaigns and lawsuits against the presidency. These forces together generated information about what the presidency was doing in its secret wars, forced it to explain its actions, and often changed the President’s desired course (and imposed significant constraints) when his explanations failed to persuade.2

The central normative claim in Power and Constraint is that contemporary checks on the presidency helped achieve national consensus — as a matter of law and policy — about the President’s proper counterterrorism powers and in the process legitimated them.3 Power and Constraint distinguishes this claim from the normative question of whether the checks produce optimal presidential constraint and optimal counterterrorism policies, issues about which the book is agnostic. There is much to admire in our constitutional system, but whether it achieves optimal counterterrorism constraints and policies depends on facts we do not (and cannot) know and on contested normative judgments.4

Professor Neal Katyal has written a thoughtful review of my book that takes issue with both its descriptive and normative arguments. I am grateful for this opportunity to respond.

I. Ex Ante Versus Ex Post Checks And Balances

Katyal claims that “traditional” or “Madisonian” checks and balances5 operate as ex ante limits on the president that “constrained [the President] at the outset instead of down the road.”6 He contrasts this supposedly traditional approach with the “newfangled checks and balances”7 described in Power and Constraint, which he says involve “ex post scrutiny”of the presidency.8 This description — both of what the Framers envisioned and the way modern checks and balances work — is inaccurate.

Katyal’s notion of ex ante checks is that each branch (most notably the Executive) should respect the constitutional and other legal limits on their power before acting, and that ex post action by the other branches should be exceptional. But neither Madison nor the framers generally believed that the President (or the other branches) would be constrained through ex ante constraints alone, or even primarily. Madison believed that the solution “for maintaining in practice the necessary partition of power among the several departments” was to “contriv[e] the interior structure of the government as that its several constituent parts may, by their mutual relations, be the means of keeping each other in their proper places.”9 As the highlighted phrases from The Federalist No. 51 demonstrate, Madison viewed checks and balances as a process of dynamic interaction between the branches.

Other passages in The Federalist No. 51 make plain that this dynamic interaction involved a heavy dose of reactive (in Katyal’s terms, ex post) measures against the usurping branch. “[T]he great security against a gradual concentration of the several powers in the same department, consists in giving to those who administer each department the necessary constitutional means and personal motives to resist encroachments of the others,” Madison noted.10 “[T]he constant aim is to divide and arrange the several offices in such a manner as that each may be a check on the other,” he added.11 And most famously, Madison said: “Ambition must be made to counteract ambition.”12 Resisting encroachments, checking other branches, counteracting ambition — all of these famous ideas suggest that Madison contemplated the branches having the interest and the power to push back against a usurping branch when it overreached. Using Katyal’s terms, Madison emphasized the need for ex post checks.

Even this way of putting things is misleading. When Katyal talks about ex post and ex ante checks, he means checks after and before presidential action, respectively. But of course the President and the other branches are always in an ex post and ex ante situation at the same time. When the President uses force abroad without congressional authorization (or takes any other potentially controversial executive action), he assesses the expected reaction of the other branches based in part on their past actions, including statutes and judicial precedents. Such actions show how the other branches reacted to similar situations in the past and how they explained what they thought would be lawful or appropriate in the future. Sometimes the President declines to act because he anticipates, based on these factors (ex ante checks), a reaction from the other branches that makes the action unattractive. And sometimes he acts and the other branches push back to defeat or limit his action (an ex post check). Madison of-ten talked in terms of ex post checks, but he surely realized that ex post checks would have ex ante effects on the presidency.13

The same point holds for what Katyal says are “Goldsmith’s ex post checks,”14 but which in reality are our nation’s system of constitutional checks that have simultaneous ex post and ex ante elements. Presidents Bush and Obama acted in the face of prior laws and precedents that informed what they believed was the scope of possible and appropriate presidential action. When their actions (such as the Bush administration’s black site and interrogation program) or threatened actions (such as the Obama administration’s threat to close the U.S. Naval Station at Guantanamo Bay) were deemed by the other branches to go too far, those branches pushed back using what in Katyal’s terms would be ex post action. But the novel forms of constraint and ac-countability that grew up after 9/11 also have ex ante effects on the presidency going forward.15

The real distinguishing feature of the modern separation of powers is not ex post checks, but rather the gargantuan array of eyeballs gazing at the presidency, in secret and in public, and forming what Power and Constraint describes as a “presidential synopticon.”16 In the presidential synopticon, many actors inside and outside the executive branch watch what it is doing, often say “no” to proposed action, force it to account for actions taken, and alter its actions if they disapprove — all with simultaneous ex ante and ex post effects.

II. Normative Claims

Katyal worries that the checks and balances described in Power and Constraint are less effective than traditional checks and balances in controlling the presidency.17 This claim is premised on the confusion about ex ante and ex post checks noted above, but also suffers from other difficulties.

The presidential synopticon was born in response to Nixon’s imperial presidency but did not come to maturity until after 9/11.18 It was first instituted because the presidency was deemed too large, too powerful, and too secretive to be monitored and held accountable by Congress and the courts acting alone. And it grew significantly in power and scope after 9/11 “because of extensive concerns about excessive presidential power in an indefinite and unusually secretive war fought among civilians, not just abroad but at home as well.”19 The reticulate system of laws that today govern national security, the literally thousands of national security lawyers who ensure compliance with these laws, the array of powerful and effectively independent inspectors general and compliance officers, the mass of reporting and related disclosure requirements, a robust press and FOIA regime, aggressive NGOs, and much more are all a response to fears of a too-big, too-potent presidency. These institutions are a supplement and spur to congressional and judicial review, not a substitute for them. Katyal suggests that the presidency would be more accountable and more constrained without these actors assisting Congress and courts.20 But he never explains why or how that is so, and it is hard to see how he might be right.

Katyal makes much of the President’s “huge first-mover advantage.”21 But the inferences he draws from this undoubtedly true premise are wrong. The framers cherished this quality in the powerful presidency they purposefully created. “Energy in the executive is a leading character in the definition of good government,” noted Alexander Hamilton in The Federalist No. 70.22 The energetic presidency can become a usurping one, of course, and that is where constitutional checks come in to play. What Katyal overlooks —surprisingly, since he argued vigorously against excessive legal constraints on the national security presidency as Deputy and Acting Solicitor General23 — is the possibility of an insufficiently energetic or overly constrained presidency, such as the ones we had on 9/11 or during the nearly successful attack by underwear bomber Umar Farouk Abdulmutallab over Detroit in December 2009.24

Katyal focuses exclusively on whether the president is adequately checked. But Congress and courts might go too far in checking the presidency, as I suspect Katyal thinks happened when Congress limited President Obama’s discretion over Guantanamo Bay detainees.25 The hard issue for normative separation of powers is not simply how to constrain the presidency, but how to ensure that the presidency and other branches are adequately restrained and empowered. A related point is that the President’s first-mover advantage does not, as Katyal thinks, inevitably lead to more presidential power. Sometimes it does. But sometimes the first move overreaches, inviting constraints that narrow presidential power from the original baseline.26

Katyal’s simplistic discussion of ex ante constraints on the presidency minimizes the difficulties of identifying and enforcing such constraints. It also overlooks the many realities about how the branches interact in practice and the contingencies that unfold without warning in a complex world. And it suffers from its failure to provide or grapple with the normative criteria for optimal presidential constraint. The identification of such criteria is among the hardest problem in normative theories about separation of powers, and is almost certainly unsolvable in general terms.27 The optimal level of presidential constraint — in national security and other contexts — is elusive because it depends on ever-changing and sometimes unknowable facts (about, for example, the nature of the threat or the efficacy of particular counterterrorism policies), as well as contested normative judgments (about, for example, what the Constitution permits or what morality requires). Katyal assumes away these difficulties but his argument depends on his ability to answer them.

Power and Constraint is agnostic about whether our separation of powers system produces such optimal constraints. It argues that the presidential synopticon pushed back against perceived extremes in the two post-9/11 presidencies and helped to achieve national consensus and legitimation. But it emphasizes, “To say that the presidential [ac-countability system] helped generate a consensus about the counterterrorism policies the President can legitimately use does not, unfortunately, mean that it generated the right policies — the ones best designed to prevent terrorist attacks while at the same time preserving other values as much as possible.”28 Optimal policies and optimal constraints are fluid, contextual, contested, and ever-changing because the facts and normative intuitions that underlie them are fluid, contextual, contested, and ever-changing.29

This fluidity of facts and norms, and the related inability to specify static optimal constraints on presidential power, are the main reasons why, as chapter 8 of Power and Constraint argues, the nation is destined to continue its cycle of under- and over-reaction to the terrorist threat. Katyal’s hopeful plea to get constraints right ex ante is unrealistic and is belied by our constitutional history.30 This sounds like a discouraging conclusion, but it should not be: “While it is difficult to make firm conclusions about optimal counterterrorism policies and their associated accountability mechanisms, the experiences of the last decade provide a second-order solace.”31 The second-order solace is that “[t]he presidential synopticon incessantly generates new information about the terror threat and the appropriateness and efficacy of counterterrorism measures to meet this threat, and our flexible political and legal institutions respond relatively quickly to this information.”32 As Arthur Schlesinger put the point, the signal virtue of our constitutional separation of powers is not that it generates perfect policies but rather that it embodies “the vital mechanism of self-correction” — the ability to generate new information about government action over time and to shape that action in ways that reflect the approval of the other institutions of government and of the American people.33 We can expect no more.

*Henry L. Shattuck Professor of Law, Harvard Law School. Since some of the matters under discussion here occurred in the George W. Bush administration, I should note that I served in that administration as Assistant Attorney General, Office of Legal Counsel (2003-2004) and as Special Counsel to the General Counsel (2002-2003). Thanks to Professor John Manning and Claire Guehenno for discussion and assistance.

1. Eric A. Posner & Adrian Vermuele, The Executive Unbound 4 (2010); see also Bruce Ackerman, The Decline and Fall of the American Republic (2010); Garry Wills, Bomb Power: The Modern Presidency and the National Security State (2010). Posner’s and Vermeule’s thesis about the legally unconstrained executive branch extends beyond the military and national security context, but that context is one of their main focuses. The other books focus primarily on national security.

2. The descriptive claims sketched in this paragraph are found in Jack Goldsmith, Power and Constraint: The Accountable Presidency After 9/11 (2012), chs. 2-6, at 23-203.

3. See id. at xv-xvi, 209-10.

4. Id. at xvi, 210-11, 251-52.

5. Neal Kumar Katyal, Stochastic Constraint, 126 Harv. L. Rev. 990, 991 (2013) (reviewing Goldsmith, supra note 2).

6. Id. at 1010.

7. Id. at 994.

8. Id. at 992.

9. The Federalist No. 51, at 317-18 (James Madison) (Clinton Rossiter ed., 2003) (emphasis added).

10. Id. at 318-19 (emphasis added).

11. Id. at 319 (emphasis added).

12. Id. (emphasis added).

13. A comprehensive theory of U.S. separation of powers, which is obviously beyond this Response’s scope, would identify primarily ex ante mechanisms (such as vetos and failures to confirm), primarily ex post mechanisms (such as the pardon power, judicial review, and congressional oversight), and the ex ante and ex post effects of their dynamic interaction.

14. Katyal, supra note 5, at 998.

15. See, e.g., Goldsmith, supra note 2, at 207, 218, 230-31, 237-39.

16. Id. at 205-43.

17. Katyal, supra note 5, at 4-14.

18. Goldsmith, supra note 2, at 34-39, 207-09.

19. Id. at 208. For additional explanation, see id. at 207-09.

20. See, e.g., Katyal, supra note 5, at 991, 1001-04.

21. Id. at 994. See generally id. at 994-1004.

22. The Federalist No. 70 (Alexander Hamilton), supra note 9, at 421.

23. See, e.g., Al Maqaleh v. Gates, 605 F.3d 84, 86 (D.C. Cir. 2010) (Katyal as counsel of record and oral advocate); see also Ashcroft v. al-Kidd, 131 S. Ct. 2074, 2078-79 (2011) (Katyal as counsel of record and oral advocate).

24. The Senate Select Committee on Intelligence identified “fourteen specific points of [executive branch] failure — a series of human errors, technical problems, systemic obstacles, analytical misjudgments, and competing priorities — which resulted in Abdulmutallab being able to travel to the United States on December 25, 2009.” S. Select Comm. on Intelligence, Unclassified Executive Summary of the Committee Report on the Attempted Terrorist Attack on Northwest Airlines Flight 253, at 2 (2010), available at

25. The restrictions began with the Department of Defense Appropriations Act, 2010, Pub. L. No. 111-118, § 9011, 123 Stat. 3409, 3466-68 (2009). The most recent restrictions are in the National Defense Authorization Act for Fiscal Year 2013, Pub. L. No. 112-239 §§ 1025, 1027—28, 126 Stat. 1632, 1913-17.

26. See generally Jack Goldsmith, The Terror Presidency (2007).

27. Elizabeth Magill’s work influenced my thinking quite a lot on this issue as I was writing Power and Constraint. See generally M. Elizabeth Magill, Beyond Powers and Branches in Separation of Powers Law, 150 U. Pa. L. Rev. 603 (2001); M. Elizabeth Magill, The Real Separation in Separation of Powers Law, 86 Va. L. Rev. 1127 (2000). For similar views, see Eric A. Posner, Balance-of-Power Arguments and the Structural Constitution (Univ. of Chi. Inst. for Law & Econ., Working Paper No. 622, 2012), available at

28. Goldsmith, supra note 2, at 210. Chapter 7 is devoted to exploring this issue.

29. Relatedly, Katyal is wrong to say that my view about possible presidential overreach is “not to worry.” Katyal, supra note 5, at 992. Power and Constraint emphasizes the need to worry about a too-powerful presidency, and describes the presidential synopticon as a response to that worry. See, e.g., Goldsmith, supra note 2, at xvi, 69-70, 208, 243, 251-52. It also notes that “[t]he continued efficacy of the presidential synopticon depends on [a] skeptical attitude about the synopticon’s efficacy.” Id. at 241.

30. I have no qualms with Katyal’s proposals for sunset laws, and indeed mentioned them as an element of congressional control in Power and Constraint. See Goldsmith, supra note 2, at 183. But they are not close to a comprehensive solution to the problem of presidential power.

31. Id. at 232.

32. Id.

33. Arthur M. Schlesinger, Jr., The Imperial Presidency 473 (First Mariner Books 2004) (1973).