Response to: Jon D. Michaels, An Enduring, Evolving Separation of Powers, 115 Colum. L. Rev. 515 (2015).


Administrative law and constitutional law enjoy a complicated relationship in legal scholarship. There is plenty of scholarship on the constitutional law of administrative law. 1 See infra notes 2–4 and accompanying text. Perhaps more than any area of law, scholars continue to debate the basic constitutionality of the entire field of administrative law. 2 See, e.g., Gary Lawson, The Rise and Rise of the Administrative State, 107 Harv. L. Rev. 1231, 1231 (1994) (“The post-New Deal administrative state is unconstitutional, and its validation by the legal system amounts to . . . a bloodless constitutional revolution.”); Cass R. Sunstein, Constitutionalism After the New Deal, 101 Harv. L. Rev. 421, 447–48 (1987) (arguing administrative state “altered the constitutional system in ways so fundamental as to suggest that something akin to a constitutional amendment had taken place”). This debate about basic constitutionality also produces doctrines constructed in light of constitutional concerns, such as the nondelegation doctrine. See, e.g., Cass R. Sunstein, Is the Clean Air Act Unconstitutional?, 98 Mich. L. Rev. 303, 330–31 (1999) (“[T]he old nondelegation doctrine . . . requires Congress to state an ‘intelligible principle’ by which to guide and limit agency action . . . [because] [i]f Congress gives the executive a ‘blank check,’ . . . it has violated Article I.”). There is also plenty of scholarship debating how constitutional doctrines like procedural due process constrain the actions of administrative actors. 3 See, e.g., Kathryn A. Watts, Rulemaking as Legislating, 103 Geo. L.J. 1003, 1046–49 (2015) (exploring procedural due process doctrine within context of administrative rulemaking). See also Goldberg v. Kelly, 397 U.S. 254, 261–62 (1970) (“[Termination of welfare benefits] involves state action that adjudicates important rights.”). It should be noted how many of the leading scholars of administrative law are also scholars of consti­tutional law. It is possible to imagine that administrative law scholars might be teachers and scholars of other doctrinal areas, but my guess is that the second-choice area of teaching and scholarship for administrative law professors tends to be constitutional law. Constitutional law dominates legal scholarship, 4 See David Fontana, The Rise and Fall of Comparative Constitutional Law in the Post-War Era, 36 Yale J. Int’l L. 1, 13 (2011) (“The attention of American legal scholars turned . . . to the Warren Court and the developments in constitutional law generated by Warren Court decisions.”). and it dominates no other area of legal scholarship more than it dominates scholarship on administrative law.

This generates many unfortunate intellectual results, but one of them is that there has been comparatively little to say about administrative law as constitutional law. Scholars have not paid sufficient attention to the inner workings of administrative law and have neglected to think through what those inner workings mean for constitutional law—rather than the other way around. In many countries, admini­strative law is constitutional law. 5 In many of the Commonwealth countries, until recently there was nothing
like judicial review constraining the political branches of government by invalidating
their actions. See Stephen Gardbaum, The New Commonwealth Model of Constitutionalism, 49 Am. J. Comp. L. 707, 708–09 (2001) (“Between 1982 and 1998 . . . [several] Commonwealth countries that were previously among the very last democratic bastions of traditional legislative supremacy . . . adopted a bill of rights . . . [that] grant[ed] courts the power to protect rights . . . .”). It was substantially by ensuring that statutes were complied with in the first place that the government was constrained by courts. See id. at 731 (noting how much of “practice and legitimacy of judicial articulation and enforcement of rights” was via “administrative law context in particular”).
If you want to understand the basic law of empowering and constraining government, you look to administrative law. Individual rights are adjudicated as “human rights,” but the questions about structural constitutional law that feature in American constitutional law feature in administrative law discussions in many countries. This is particularly true in the Commonwealth countries, where there simply was no constitutional law until recently. 6 See id. at 708–09.

These debates about administrative law as constitutional law have not transpired on the American scholarly terrain, with a few exceptions. One such exception in the past few years is the important scholarship produced by Professor Jon Michaels. In a series of important articles, and now a forthcoming book, Michaels models administrative law as constitutional law. 7 See Jon D. Michaels, The Second Privatization Revolution: The Unmaking of the American State (forthcoming 2016); Jon D. Michaels, All the President’s Spies: Private-Public Intelligence Gathering in the War on Terror, 96 Calif. L. Rev. 901 (2008); Jon D. Michaels, Deputizing Homeland Security, 88 Tex. L. Rev. 1435 (2010); Jon D. Michaels, Of Constitutional Custodians and Regulatory Rivals: An Account of the Old and New Separation of Powers, 91 N.Y.U. L. Rev. (forthcoming 2016); Jon D. Michaels, Privatization’s Pretensions, 77 U. Chi. L. Rev. 717 (2013); Jon D. Michaels, Privatization’s Progeny, 101 Geo. L.J. 1023 (2013); Jon D. Michaels, Running Government Like a Business . . . Then and Now, 128 Harv. L. Rev. 1152 (2015) (reviewing Nicholas R. Parrillo, Against the Profit Motive: The Salary Revolution in American Government, 1780–1940 (2013)); Jon D. Michaels, The (Willingly) Fettered Executive: Presidential Spinoffs in National Security Domains and Beyond, 97 Va. L. Rev. 801 (2011). Michaels is foregrounding the claim that the administrative state is giving us now much of what constitutional law has given us historically.

In his latest article in this area, Michaels makes essentially two arguments. First, the administrative separation of powers between political appointees, civil servants, and civil society have “dispositional characteristics” similar to the three branches of government in constrain­ing one another. 8 See Jon D. Michaels, An Enduring, Evolving Separation of Powers, 115 Colum. L. Rev. 515, 525, 530 (2015) [hereinafter Michaels, Separation of Powers]. Political appointees in administrative agencies play the role of the President in the older three-branch model of federal power. 9 See id. at 538–40. The civil service plays the role of the objective and technocratic judicial branch. 10 See id. at 540–47. Civil society plays the role of the broadly representative and deliberative legislative branch. 11 See id. at 547–51.

Second, Michaels argues that the role each of the three actors plays in the administrative separation of powers is jeopardized by the rise of increasingly significant and unaccountable privatized administrative power. 12 See id. at 570–96 (“[P]rivatization’s commingling and consolidation of political and commercial power endangers administrative separation of powers . . . .”). We transitioned from the constitutional separation of powers to the administrative separation of powers. 13 See id. at 529–70. We might now be transitioning to the “privatized era.” 14 See id. at 570–96.

This Essay will address the first of these two arguments. Part I argues that any claim about separation of powers must demonstrate that separated powers are different powers. In other words, each separate power must “resist encroachments” 15 The Federalist No. 51, at 289–90 (James Madison) (Clinton Rossiter ed., 1961). and “enforce limitations on [the other] powers.” 16 Akhil Reed Amar, Of Sovereignty and Federalism, 96 Yale L.J. 1425, 1495 (1987). Accounts of how separation of powers fails often argue that there was not enough difference to generate enough separation. 17 See, e.g., Daryl J. Levinson & Richard H. Pildes, Separation of Parties, Not Powers, 119 Harv. L. Rev. 2311, 2315 (2006) (“[T]he degree and kind of competition between legislative and executive branches . . . may all but disappear, depending on whether the House, Senate, and presidency are divided or unified by political party.”).

Part II offers a question for Michaels’s claim. If we want to be sure that institutional actors are sufficiently different, are we sure that the three stakeholders in the administrative state meet that requirement? Given that each one of the three actors in his separation of powers is composed mostly of Washington political (and often legal) elites, do we really believe that there is enough difference of powers to generate enough separation?

I. The Separation of Powers as the Difference of Powers

The original constitutional design was very conscious of the fact that “a mere demarcation on parchment of the constitutional limits of the several departments is not a sufficient guard against those encroach­ments which lead to . . . government in the same hands.” 18 The Federalist No. 48, supra note 15, at 308 (James Madison). Constitutional designs had to generate institutions that were adequately motivated to act sufficiently differently. A difference of powers generates many of the institutional goods that separation of powers was meant to provide, from representation to efficacy to checks and balances. 19 See David Fontana, The Geographical Separation of Powers (forthcoming 2016) (noting how separation of powers presupposes institutional differentiation). Michaels’s article references the need for a difference of powers, 20 See Michaels, Separation of Powers, supra note 8, at 525 (“The Framers . . . endowed each group with distinct dispositional, political, and institutional characteristics . . . . And they made each group answerable to different sets of constituencies . . . .”); see also David E. Pozen, Self-Help and the Separation of Powers, 124 Yale L.J. 2, 75–76 (2014). and a discussion of this prerequisite for separation of powers can be found in many places in The Federalist Papers. 21 See, e.g., The Federalist No. 47, supra note 15, at 302–03 (Alexander Hamilton) (noting importance of “opposite and rival interests”); The Federalist No. 10, supra note 15, at 63–64 (James Madison) (noting importance of “distinct parties and interests”); The Federalist No. 51, supra note 15, at 320 (James Madison) (noting importance of different “wills” and “ambitions”).

Designs producing difference essentially happen as selection effects 22 See Adrian Vermeule, Selection Effects in Constitutional Law, 91 Va. L. Rev. 953, 953 (2005) (“Constitutional rules . . . should focus not only on the creation of optimal incentives for those who happen to occupy official posts at any given time, but also on the question of which (potential) officials are selected to occupy those posts over time.”). Institutional differentiation must feature mechanisms that select for different types of individuals and different type of ideas for each institution. and as treatment effects. 23 Professor Adrian Vermeule calls these “incentive-based” accounts of institutional design. Id. at 953 (“The literature on constitutional design focuses on the incentives that shape the behavior of government officials and other constitutional actors.”). Institutional differentiation must also feature mechanisms that ensure—once selected—that those in each branch of government are exposed to different forces shaping their behavior. In the original design, four different selection mechanisms (one for the House, one for the Senate, one for the President, and one for the courts) provide the three branches with their respective powers. For the elected branches, separate electoral constituencies manufacture difference. Individual districts in each state elect the House of Representatives. 24 See U.S. Const. art. I, § 2. The Senate is elected by entire states. 25 See id. art. I, § 3, cl. 1, amended by U.S. Const. amend. XVII. The President is elected by the entire nation. 26 See id. art. II, § 1. For the judicial branch, federal judges are selected by presidential nominations and senatorial advice and consent. 27 See id. art. III, § 1. The branches also have different tenures. While the President is selected every four years and federal judges are appointed to lifetime terms upon a vacancy, the House is selected every two years and one-third of the Senate is selected every six years. Because priorities and preferences change over time, 28 See R. Douglas Arnold, The Logic of Congressional Action 193–223 (1990) (noting just as policy preferences change over time, so do policy agenda preferences); John W. Kingdon, Agendas, Alternatives, and Public Policies 12–13 (2d ed. 1995) (same). so too will the members that are elected at different times or the judges or Justices nominated at different times.

In the original design, three different institutions faced three different incentives or “treatment effects” once they had been selected for their offices. At the level of the institution, just as separate times and separate places of selection ensure difference, there are constitutional rules to prevent collusion. No official can serve in more than one branch at a time. 29 See U.S. Const. art. I., § 6, cl. 2 (“[N]o Person holding any Office under the United States . . . shall be a member of either House.”). There are doctrines of executive privilege and legislative privilege permitting one institution to separate out its information from another actor. 30 See, e.g. William P. Marshall, The Limits on Congress’s Authority to Investigate the President, U. Ill. L. Rev. 781, 785–86 (2004) (summarizing these doctrines). Our system creates different powers for different branches. Article I is replete with powers that Congress has, such as regu­lating “Commerce . . . among the several States.” 31 U.S. Const. art. I, § 8; see also Nat’l Fed’n of Indep. Bus. v. Sebelius, 132 S. Ct. 2566, 2580 (2012) (noting importance of institutions having “constitutional authority to . . . act”). Article II specifies the powers of the President. 32 U.S. Const. art. II. Article III references the “judicial power” that the federal courts exercise. 33 Id. art. III, § 1. Scholars disagree about how to define these powers, with formalists on one side and functionalists on the other. 34 See M. Elizabeth Magill, Beyond Powers and Branches in Separation of Powers Law, 150 U. Pa. L. Rev. 603, 606 (2001) (summarizing and critiquing debate between formalists and functionalists). Yet scholars and courts agree that defining separate powers is a key tool to achieving separation. For there to be separateness, there must be some institutions doing some things, and other institutions doing other things.

II. Are Administrative Stakeholders Different?

The question for the argument that the administrative separation of powers replicates or at least approximates the constitutional separation of powers, then, is whether the administrative separation of powers produces differences of this magnitude. My ambition in this Part is to raise this question, and to raise some additional questions as to why one might wonder if there will be as much differentiation in the admini­strative separation of powers as there has been in the constitutional separation of powers. Simply put, are the three branches of administrative power too similar? For each of the three parts of the administrative separation of powers, this Essay briefly argues, large numbers of the people that make up these institutions have come from similar places and aspire to go to similar places.

Let’s take, first, the politically appointed agency leaders standing in for the President. The President is elected by the entire nation once, maybe twice, and has strong incentives to be accountable to large swaths of the country while in office. 35 See Christopher C. DeMuth & Douglas H. Ginsburg, White House Review of Agency Rulemaking, 99 Harv. L. Rev. 1075, 1075 (1986) (providing classic account of this claim). The selection and treatment effects for political appointees are different—and, as we shall see, similar to the selection and treatment effects for other administrative actors. Politically appointed agency leaders tend to be highly educated elites—often lawyers—who receive their positions because of their connections within the national party or within the national presidential campaign of the President appointing them. 36 See David Fontana, The Permanent and Presidential Transition Models of Political Party Policy Leadership, 103 Nw. U. L. Rev. 1993, 1995 (2009) (noting how much of policy leadership changes as result of presidential campaigns and elections).

Political appointments are just temporary positions, with appointees usually serving just about two years. 37 See Gen. Accounting Office, GAO GDD-94-115, Fact Sheet, Political Appointees, Turnover Rates in Executive Schedule Positions Requiring Senate Confirmation 2–3 (Apr. 1994). The most comprehensive empirical scholarship on political appointees remains the work of Professor David Lewis. See generally David E. Lewis, The Politics of Presidential Appointments (2008) (providing wealth of empirical analysis and information about political appointees). This means that political appointees spend the majority of their professional career elsewhere. These earlier professional positions can be significant in shaping their later behavior as political appointees. It is a basic tenet of organizational behavior that one “does not live for months or years in a particular position in an organization, exposed to some streams of communication, shielded from others, without the most profound effects upon what he knows, believes, attends to, hopes, wishes, emphasizes, fears, and proposes.” 38 Herbert A. Simon, Administrative Behavior, at xvi (3d ed. 1976) (noting impact organizational context has on individual behaviors). It is quite common for political appointees to have spent long periods of their earlier careers in relevant civil society organizations or even in the civil service. 39 This reality is perhaps more common in some areas. Take national security policy, for instance. John Brennan was a civil servant in the Central Intelligence Agency (CIA) for twenty-five years before becoming the politically appointed head of the CIA. David A. Graham, Meet John Brennan, Obama’s Drone Czar and Nominee for CIA Director, Atlantic (Jan. 7, 2013), (on file with the Columbia Law Review). If political appointees are meant to be different, they are less different because they are socialized by the institutions that they are meant to be different from later in their careers. 40 The same is true of the private sector interacting with administrative actors in Washington. Much of the value that a private-sector employee brings to their position in Washington is their position within the same personal and professional networks as those working for administrative entities. See David Fontana, The Narrowing of Federal Power by the American Political Capital, 23 Wm. & Mary Bill Rts. J. 733, 741 (2015) [hereinafter Fontana, Narrowing of Federal Power].

We also know from the empirical work on political appointees that their exit options are significant in explaining parts of their behavior. 41 See Anthony M. Bertelli & David E. Lewis, Policy Influence, Agency-Specific Expertise, and Exit in the Federal Service, 23 J. Pub. Admin. Res. & Theory 223 (2013) (documenting how nature and desirability of exit options affect civil-service behavior and tenure). With just a short time as a political appointee, something must come next. Scholars have written about “burrowing,” the process by which former political appointees turn themselves into civil servants. 42 See, e.g., Paul C. Light, Thickening Government: Federal Hierarchy and the Diffusion of Accountability (1995); Nina A. Mendelson, Agency Burrowing: Entrenching Policies and Personnel Before a New President Arrives, 78 N.Y.U. L. Rev. 557, 606–16 (2003) (describing personnel burrowing preceding presidential transitions). It is perhaps even more common for a former political appointee to transition to civil society. Civil society organizations desiring to accomplish their policy objectives will value immensely the connections and experience that a political appointee has. This leads to one basic question for the administrative separation of powers as the administrative difference of powers: How differently will a political appointee act in the morning if in the afternoon they are interviewing with a civil service or a civil society official whose lives they could be dramatically affecting in the morning?

Let’s next turn to the institutional replica of the judiciary in the administrative separation of powers: the objective, professionalized, and technocratic civil service. Federal judges are selected through a combination of political connections and professional qualifications. Once on the bench, their incentives are many, including to impress other elites 43 See Lawrence Baum & Neal Devins, Why the Supreme Court Cares About Elites, Not the American People, 98 Geo. L.J. 1515, 1570–74 (2010) (arguing and documenting how legal elites—particularly Supreme Court Justices—are concerned with what other elites are thinking). and to further impress political leaders so as to be elevated to a higher court. 44 See Carlos Berdejo & Daniel L. Chen, Priming Ideology? Electoral Cycles Without Electoral Incentives Among U.S. Judges 27–28 (2013) (unpublished manuscript) (on file with the Columbia Law Review) (empirically documenting how lower court behavior changes because of possibility of higher-court appointment). Civil servants are highly educated elites—often lawyers—who receive their connections likewise because of powerful personal and professional networks.

Before becoming federal officials, civil servants have often spent their key moments of professional socialization in an ideologically affiliated civil-society organization. Consider, for instance, that the lawyers hired to work in the Civil Rights Division of the Department of Justice (DOJ) during the presidency of Barack Obama had worked for many civil-society organizations like the American Civil Liberties Union. 45 See Charlie Savage, In Shift, Justice Department is Hiring Lawyers with Civil Rights Backgrounds, N.Y. Times (May 31, 2011),
01/us/politics/01rights.html (on file with the Columbia Law Review) (explaining shift in hiring backgrounds of lawyers from one presidential administration to the next).
New hires during the presidency of George W. Bush had also worked in civil-society organizations—but conservative-leaning ones. 46 See id. Some civil servants may be former political appointees who burrowed into a permanent position. Indeed, some civil servants might actually be de facto political appointees, given the implicit and (sometimes) explicit professional qualifications required for some civil-service positions. 47 See David Fontana, Executive Branch Legalisms, 126 Harv. L. Rev. Forum 21, 29–30 (2013) [hereinafter Fontana, Legalisms] (arguing documentation of number of poli­tical appointees in academic literature understates how many civil servants are appointed using political considerations).

Once in office, civil servants certainly do face the incentives that Michaels identifies to do a technically proficient job because of the various forms of protection civil servants enjoy from politically-motivated retaliation. 48 See Michaels, Separation of Powers, supra note 8, at 540–42. Indeed, one study found that the average civil service employee only has a 0.03% chance of being fired in a given year. 49 Robert M. O’Neil, The Rights of Public Employees 122–23 (2d ed. 1993). Civil servants, though, face other treatment effects in office that make them less likely to be different than political appointees or civil-society personnel. The sheer number of political appointees—more than 8,000 by one account 50 See William S. Dietrich, In the Shadow of the Rising Sun 185 (1991). —means that for a civil servant to be effective, they will have to convince and perhaps replicate their political-appointee bosses. 51 For all of the empirical studies about the neutrality of civil servants that Michaels rightly cites, one could cite perhaps an equal number of studies showing the political context shaping civil-servant behavior. See, e.g., Joel D. Aberbach, Robert D. Putnam & Bert A. Rockman, Bureaucrats and Politicians in Western Democracies 94–95 (1981) (comparing more political self-identity of American civil servants with more technocratic self-identity of civil servants elsewhere); Bruce Ackerman, The New Separation of Powers, 113 Harv. L. Rev. 633, 702–10 (2000) (noting incentives civil servants face to act politically); B. Dan Wood & Richard W. Waterman, The Dynamics of Political Control of the Bureaucracy, 85 Am. Pol. Sci. Rev. 801, 802 (1991) (providing empirical evidence of political influences on civil-servant behavior). Many civil servants will aspire to be a (more powerful) political appointee. 52 See Fontana, Legalisms, supra note 47, at 31. This is particularly true for the first layer of civil servants, who often receive their positions because of similar networks to those that help political appointees. 53 See id. at 29–30 (noting how many civil servants receive their positions in process similar to how political appointees receive their positions). Many civil servants will also transition to civil society, using their expertise at DOJ Civil Rights to go work for the legal office of the NAACP, for instance. This all leads to the same question: How different are civil servants from political appointees if they come from the same place and aspire to go to the same place?

Finally, let’s take civil society. In the administrative separation of powers, civil society is meant to stand in for Congress. Like Congress, civil society is meant to be a large and inclusive institutional actor. Congress is selected by all kinds of people all over the country and serves all kinds of people all over the country, including narrower categories of people than the broad national electorate that presidents tend to be responsive to. Civil society is meant to do the same.

The question, again, is whether civil society features too many Washington elites to qualify as sufficiently different from its institutional counterparts—and enough to qualify as sufficiently similar to Congress. Civil-society organizations with the resources and skills to shape administrative behavior will largely be Washington organizations, or at least the Washington offices of larger organizations. 54 See Fontana, Narrowing of Federal Power, supra note 40, at 741 (noting geographical concentration of organizations with political influence). Because of this Washington location—and because their ambition is to shape administrative behavior—civil-society organizations very much desire those with administrative experience. A political appointee can bring the connections and the competence that a civil-society organization desires. A civil servant can perhaps do some fraction of that. 55 For a discussion of the labor markets for political appointees, and why and what they do after leaving office, see Bertelli & Lewis, supra note 41. Political appointees, and to a lesser degree civil servants, will have to find something else to do eventually anyway, either because they want a change of scenery or because a President of their party is no longer in power.

Once in their position in their civil-society organization, a series of incentives present themselves. Supporting rather than battling political appointees and civil servants is more likely to yield the tangible outcomes that will please the board of directors of the civil-society organization. Civil-society funding is always fractious and fragile, and not causing too many problems for political appointees and civil servants can perhaps lead to employment opportunities with them. Even if one’s true passion is to work in the civil-society sector, finding a way to work in the administrative branch for some period of time is a substantial profes­sional attribute.


The separation of powers discussion faces a constant, essentially irresolvable question: What precise kind of separation of powers do we want? 56 See Eric A. Posner, Balance-of-Powers Arguments and the Structural Constitution (2012) (Univ. of Chi., Inst. for Law & Econ., Working Paper No. 622, 2012) (noting how much of these arguments depends on how much balance is desired). Countries that pledged their fidelity to parliamentary systems without our three-branch separation long ago abandoned the ambition of no separation. Even parliamentary systems feature separate bureau­cracies and courts. 57 See Ackerman, supra note 51, at 664–71 (noting convergence of parliamentary systems toward model of independent constraints on political branches). We are all separation of powers supporters now.

The question that really matters, then, is what type of separationist we are, not whether we are in the first place. Without a definition of how much separation is desired and for what purpose, it is more difficult to assess whether some institutional design will achieve that separation. As Professor Cristina Rodríguez raises in her thoughtful response to the Michaels article—and is true of most of the separation of powers literature—we need some sort of preexisting theory of what separation is doing before we can assess what means can get us that separation. 58 See Cristina M. Rodríguez, Complexity as Constraint, 115 Colum. L. Rev. Sidebar 179, 184 (2016) (noting how this area of scholarship does not “spend[] much time articulating and analyzing the purposes separating powers is supposed to serve”). We need to know more about ends before we can fully analyze means.

Once Michaels calibrates the degree of separation desired, he faces a central question about the precise means to achieve those separationist ends. The question posed for Michaels in this Essay about his separation is one posed (ironically) by the former Supreme Court clerk, Bush political appointee, and current U.S. Senator, Ted Cruz. How much of the administrative separation of powers is just government by the “Washington cartel?” 59 See Elise Viebeck, ‘Washington Cartel’: The Ups and Downs of a Cruz Catchphrase, Wash. Post (Aug. 12, 2015),
post/wp/2015/08/12/washington-cartel-the-ups-and-downs-of-a-cruz-catchphrase/ [http:/
Given who exercises administrative power and what their incentives are when they do, does the administrative branch really feature enough difference to produce enough separation? If Michaels wants the same degree of separation that James Madison wanted, that leads to another fundamental question: Are their means of achieving separation equally effective to unite them as a positive and normative matter?