Complexity as Constraint

Complexity as Constraint

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


In his ambitious and panoramic article for the Columbia Law Review, Jon Michaels sets out to modernize our conceptions of the separation of powers, explaining how the basic three-branch model that every American school child learns to accept on faith has weathered the test of time. In An Enduring, Evolving Separation of Powers, 1 Jon D. Michaels, An Enduring, Evolving Separation of Powers, 115 Colum. L. Rev. 515 (2015). Michaels makes two impres­sive and timely contributions to administrative law and constitutional theory. First, he provides a wide-ranging and textured account of systemic adaptation, showing how we have maintained our constitutional commit­ments to “constrained” and “rivalrous” government, despite sea changes in the scale, structure, responsibilities, and capacities of the insti­tutions that make up the three branches of government—the executive branch in particular. 2 Id. at 532–70. Second, he sounds the alarm nonetheless, sizing up a new and growing challenge to those commitments—the privatization of gov­ernment—while providing constitutional actors with some conceptual tools to confront it. 3 Id. at 570–96.

In Part I of this Essay, I begin by summarizing Michaels’s principal claims and highlighting what I believe to be his most important contribu­tions to the literature that seeks to understand the operation and reach of executive power in today’s world. On my read, Michaels successfully puts to rest concerns about an overweening executive branch. If any­thing, under his view, our system may require more, not less, government power, but with ongoing attention to the types of power that drive the reg­ulatory state.

In Parts II and III, I then offer two critiques. The first points to the limited utility of comprehensive structural theory in constitutional law. The second offers an alternative, though less reassuring, account of what keeps government constrained. On the first score, despite Michaels’s con­siderable effort to demonstrate how constrained, rivalrous govern­ment has persisted to control power within the modern administrative state, it re­mains unclear precisely what he thinks ought to be con­strained. In Part II, I explore the possibilities. Is it government power and regulation gen­erally, in which case the trend toward privatization looks like a happier story than the one he presents? Is it overweening political control of the bureaucracy? Or is it the unaccountable exercise of power? The answer may well depend on time, place, and regulatory setting, which will make a coherent theory of the separation of powers elusive. But to identify whether the system requires an adaptation or recalibra­tion, we must have some sense of what constitutes the abuse of power.

Second, even if we cannot come to agreement in any general sense on the purposes of constraint, we can still count on executive power be­ing held in check, but without worrying about the particular mix of pow­ers Michaels describes as necessary. I argue in Part III that something far simpler constrains the President and the executive branch today—the complexity of government. Both the scale of the government and the di­versity of interests within it work against the concentration of power. This form of constraint might be of limited comfort for those who believe par­ticular institutions should have specific responsibilities that serve as pre­dictable and prophylactic devices for keeping power in check. It will also at times frustrate two other camps whose interests are them­selves often in tension with one another—those who regard the separa­tion of powers as a means of reducing the government’s footprint in our lives as a whole and those who lament obstacles to the government’s abil­ity to regulate to solve social problems. But this complexity does work against the concen­tration of power, and in that sense, it means we have less to fear from executive-driven government than standard accounts predict.

I. The Internal Separation of Powers

In his article, Michaels describes three moments in the history of the separation of powers, each of which has been marked by concern for out­sized power, albeit emanating from different sources. Whereas the Founders worried about “the prospects of an imperious, domineering national legislature,” 4 Id. at 517. the central concern of the twentieth century was “the executive-driven administrative juggernaut of the New Deal.” 5 Id. Chan­neling his own research preoccupations, he then says that “[t]oday, we worry about the rampant commingling and concentration of state and commercial power associated with privatization.” 6 Id. at 517–18. Finally, Michaels issues a call to arms. Just as scholars and officials in each of those previous eras created institutions and generated norms and practices to hold ascend­ant government power in check, “the time is . . . right to anticipate, chal­lenge, and again most significantly start to constrain . . . burgeoning political-commercial partnerships.” 7 Id. at 519–20.

Michaels’s dissection of the privatization phenomenon and his effort to link its rise to familiar separation of powers debates augments the vital contributions of his body of scholarly work generally. By showing how moves toward privatization degrade the civil service and enlist certain members of the public in the expansion rather than the constraint of regulatory authority, 8 See id. at 577 (describing how privatization serves to “unshackle” state). he links new trends in governance to time-honored constitutional precepts. But as Michaels himself recognizes, it can be dif­ficult to know how pervasive and lasting privatization might be, and to declare a Privatization Era might be premature. Without question, these new forms of doing business demand scrutiny to see if they compromise systemic values. But it remains too early to tell if privatization will trans­form government on the order of the modern administrative state and therefore demand empowerment of a new constituency to check privat­ization’s rise. Just as likely, the potential pathologies of privatization re­flect only a variation on the challenges of the administrative era, requir­ing solutions familiar within the twentieth-century narrative Michaels provides. It may well be that challenges of privatization can be met from inside the administrative separation of powers through context-specific tweaks to government contracting regimes and the other sources of pri­vatization he identifies. 9 Another possibility is that the rise of privatization has been a product of a general distrust of government that has fueled the Republican Party and related political and juris­prudential movements. This distrust informs growing preferences for private actors to per­form formerly public functions. It also animates the reluctance to raise taxes to fund govern­ment programs and leads to the discrediting of, or disrespect for, the sort of civil service that would be required to perform the professionalized checking functions Michaels as­cribes to it. Addressing these sorts of dynamics seems like a deep cultural and political challenge, not a matter for constitutional law or even administrative design.

Michaels emphasizes that his efforts at periodization do not mean that the dynamics of previous eras no longer persist today. 10 Michaels, supra note 1, at 523. After all, Congress still exists, and debates about its power relative to the presi­dency and the courts—the classic separation questions—go on. 11 See, e.g., Free Enter. Fund v. Pub. Co. Accounting Oversight Bd., 561 U.S. 477, 495–97 (2010) (invalidating provision of Sarbanes-Oxley Act creating dual for-cause re­moval structure and concluding that obstructing President’s removal power in this way inter­fered in vesting of executive power in President). But Michaels’ construction of a Privatization Era runs the risk of obscuring that we remain deeply embroiled in the debates of the administrative era. He identifies the concern over executive power as ascendant in the 1930s and 1940s, but the debates that have produced the administrative separa­tion of powers 12 See Michaels, supra note 1, at 526–29 (noting “modern, federal administrative state really took off in the 1930s and 1940s,” which “signaled that the Executive was now the constitutional institution to reckon with”). remain central to how we understand the risks and bene­fits of government power. Our post–New Deal reality is one of pervasive and capacious executive power. 13 See, e.g., Eric Posner & Adrian Vermeule, The Executive Unbound 3–18 (2011) (persuasively describing extent and dominance of executive power, particularly in circum­stances of crisis). The rise of the national security and surveillance state, which accelerated after September 11, 2001, has only deepened and broadened the executive branch’s reach, as well as its ambitions.

Accordingly, the most important contributions of his article may be to our understanding of the very administrative era he says is under threat. I therefore focus primarily in this Essay on what Michaels has to say about the administrative separation of powers. His account effectively deflates ongoing concerns about an overweening Executive, but in a way that departs from familiar claims. 14 Compare Bruce Ackerman, The Decline and Fall of the American Republic 181 (2010) (arguing executive power has spiraled dangerously out of control and existing insti­tutional arrangements no longer constrain President, though with emphasis on foreign af­fairs and war powers), with Posner & Vermeule, supra note 13, at 113–43 (arguing law does not constrain Executive but politics does). He demonstrates that our legal cul­ture has endowed new institutions with the “dispositional characteris­tics” 15 Michaels, supra note 1, at 530. of the classic branches, providing the tension in government nec­essary to keep it constrained. Politically accountable agency leaders stand in for the politically motivated and quick-to-act executive branch. The civil service provides the sober, disinterested perspective of the judiciary. And civil society, empowered through the Administrative Procedure Act and the imperatives of regulatory governance, provides the wide-ranging, deliberative input of the legislature.

This work is of a piece with and clearly informed by the growing body of scholarship focused on the institutional design of the executive branch and the separation of powers and functions within it. 16 Adam B. Cox & Cristina M. Rodríguez, The President and Immigration Law Redux, 125 Yale L.J. 102, 197 (2015) (arguing we should understand President Obama’s efforts to extend relief from deportation to unauthorized immigrants as efforts to channel and con­trol executive power from within); Neal Kumar Katyal, Toward Internal Separation of Powers, 116 Yale L.J. Pocket Part 106, 106–10 (2006),
ernal-separation-of-powers [] (arguing for implementation of sep­aration of powers principles within executive branch given President’s vast power); Elizabeth Magill & Adrian Vermeule, Allocating Power Within Agencies, 120 Yale L.J. 1032, 1078 (2011) (discussing how allocation of responsibilities within agencies can affect agency’s functions and outputs by empowering different sorts of constituencies); Gillian E. Metzger, The Interdependent Relationships Between Internal and External Separation of Powers, 59 Emory L.J. 423, 425 (2009) (drawing attention to internal admin­istrative design and evaluat­ing different structures for extent to which they serve as effec­tive checks on government).
Whereas standard administrative law questions ask when and how the traditional branches—Congress and the courts—intervene to check agencies’ power, 17 Michaels, supra note 1, at 532. Michaels and his fellow travelers recognize these “overseers” as “fickle” and “easily distracted.” 18 Id. at 533. They focus, instead, on the inner workings of the Executive itself, acknowledging that “durable consistent checks [on government actually] operate on the ground.” 19 Id.

The rise of this scholarship makes a great deal of sense as a matter of academic sociology and intellectual history. It reflects both dismay and resignation over the decline of Congress as a robust deliberative institu­tion during an era of hyperpolarization. But rather than characterize Congress’s demise as enabling the rise of executive tyranny, this work highlights that executive-driven governance has been and can be kept in check. It may be that we would be better off as a polity with a more active Congress capable of reaching compromise across party lines to legislate with greater frequency. But the goal of constrained and rivalrous govern­ment remains attainable, even as much of the activity of the executive branch occurs beyond Congress’s capacity to supervise.

In many of the byways of his article, Michaels adds nuance to what we already have come to appreciate about this internal separation of powers. But as he emphasizes, he brings to this scholarly agenda an ap­preciation for how constraints on power do not just arise from within, as a function of how Congress has designed a particular agency or regula­tory domain. 20 See id. at 536–37 (“[M]y tripartite scheme stands out as a framing device . . . be­cause the interplay among agency leaders, civil servants, and civil society is the common, trans­substantive denominator in administrative governance . . . .”). He captures as well the external dimensions of a rivalrous separation of powers, namely the role that civil society plays in supervis­ing government. 21 Id. at 536. For other accounts of external forms of influence, see Jack Goldsmith, Power and Constraint: The Accountable Presidency After 9/11, at 51–82 (2012) (explor­ing how media and advocacy by nongovernmental organizations (NGOs) have constrained ex­ecutive branch); Posner & Vermeule, supra note 13, at 113–53 (arguing politics punish and prevent execu­tive abuses). He provides us, then, with a capacious understanding of accountability, showing how it depends not only on internal vigilance but also on a mobilized populace, which can consist of interest groups, concerned citizens, self-appointed private watchdogs, and even market interests. 22 See Michaels, supra note 1, at 547–51 (“[E]mpowered and often highly motivated members of civil society use administrative procedures to educate and hold agency leaders (and civil servants) accountable, limiting opportunities for those officials to proceed arbi­trarily, capriciously, or abusively.”).

Even more intriguing than this broadening of perspective, though, is what Michaels’s concept of administrative separation suggests about power generally. His self-conscious association of the internal separation of powers with the three pillars of the Madisonian regime helps to reha­bilitate and transform the idea that different types of power exist and ought to be intermingled with one another in any system of democratic governance. 23 See Jeremy Waldron, Separation of Powers in Thought and Practice?, 54 B.C. L. Rev. 433, 438, 442, 467 (2013) (arguing principle of separation of powers, even if “dying,” “still . . . points to something that was once deemed valuable—namely, articulated govern­ment through successive phases of governance each of which maintains its own integrity”). Strict adherence to the Madisonian framework would re­quire that such power be located in discrete branches of government. But as Michaels implicitly recognizes, we are long past a time when it made sense to resist a blending of functions among the branches—to ar­gue that executive actors may not adjudicate, or legislate, for exam­ple. 24 See M. Elizabeth Magill, Beyond Powers and Branches in Separation of Powers Laws, 150 U. Pa. L. Rev. 603, 605–06 (2001) (demolishing this concern with formal separa­tion and emphasizing how governmental power is shared by “large and diverse set of gov­ernment decisionmakers”). Critics with a strong attachment to Madisonian separation might still claim that an internal distribution of different powers within the ex­ecutive branch will not do. Despite its size, they might argue, the Executive remains uni­tary in the sense of being headed by a single figure and his inner circle of advisors and appointees. Neither a civil service located within that branch nor a highly diffuse civil society outside it will provide counterweights of the necessary magnitude. For that, we require the full-blown institutions of Madisonian separation designed to engen­der rivalries. But Michaels’s story is one of adaptation—of an implicit acceptance of the modern realities of government and a search for per­haps second-best counter­weights to the form of power (executive power) that modern realities have made dominant.

II. Rivalries to What End?

The jurisprudence and academic literature concerning the separa­tion of powers share at least one surprising limitation. Neither spends much time articulating and analyzing the purposes separating powers is supposed to serve. The Supreme Court, when determining how the Constitution allocates particular powers, speaks in platitudes, extolling separation as a bulwark against tyranny. But as a co-author and I have noted elsewhere, this idea is hopelessly abstract. 25 Cox & Rodríguez, supra note 16, at 165–66. Concrete power strug­gles rarely involve a clear choice between tyranny and freedom. 26 Id. Con­stitutional formalists would say we must adhere to the structures estab­lished by the Constitution, because they are the blueprint we have for ensuring that government does not tyrannize us. But some of the most difficult separation of powers controversies cannot be resolved textually. As Michaels so ably demonstrates, the administrative era has demanded an adaptation of the Constitution’s formal structures to maintain the doc­ument’s relevance to the demands of a modern economy and welfare state. 27 Michaels, supra note 1, at 530.

Like many separation-of-powers scholars, Michaels appears less inter­ested in defining separation’s overarching purposes in order to reach proper conclusions about institutional design than in maintaining the commitment to separation itself. Throughout his piece, Michaels priori­tizes the maintenance of rivalrous government, and his work suggests that maintaining structures in tension with one another will advance the purposes of the separation of powers, even if subconstitutional actors ultimately stand in for the three branches as the relevant rivals. 28 See id. at 562–63 (describing Court’s willingness to permit interbranch interfer­ence among constitutional actors as “suggestive of the fact that the federal constitutional system insists upon assurances of encumbered, limited government, but is willing to allow subconstitutional actors to furnish those encumbrances”). But what are those purposes that our governing structures ought to serve in their competition with one another? In Michaels’s story of adaptation, what is it that we must develop new norms and institutions to preserve? Michaels does not provide us with the resources to answer these ques­tions, and in this Part, I explore why these questions matter.

Michaels seems to worry mostly that power will become concen­trated, which can then result in the government acting in an arbitrary or abusive fashion. 29 See id. at 567–70 (arguing while “there is little threat of a military coup, there re­mains reason to worry about the everyday power of the State (and state apparatchiks) to act arbitrarily and abusively in countless ways, often at the touch of a button”). He aug­ments this understanding with the claim that constraining power has the effect of legiti­mizing it. Id. Though he does not say so explicitly, the links he draws between the administra­tive sep­ara­tion of powers and the institutional pillars of Madisonianism burnish that legiti­macy by ty­ing our current structures of government to an original Founding design em­bodied in our governing charter of the Constitution. At times it seems as if Michaels believes the exercise of government power should be difficult (but not too difficult). So each type of power he identifies as integral to separation ought to be represented, and as long as each acts in good faith, the work of government will get done, but in a manner that comports with ba­sic notions of fairness and respect for constitutional rights. But there are countless ways to divide up government to prevent concentration. Even within the specific contours of the U.S. system, if we accept Michaels’s account of the three types of power it calls for, numerous possible relationships among those types of power could reasonably be said to work against concentration. To understand how to assess and calibrate these competing forms of power, we need more than a stand against concentration. We need a definition of abuse.

We can imagine at least two major sorts of abuses, which happen to be in some tension with one another, further complicating the picture. We might worry that a President will impose his ideas and preferences all the way down through the executive branch. If this were our concern, we would prioritize shoring up the civil service and perhaps reducing the number of political appointees within the system. But what exactly is the problem with such political control? Does it suggest that government is being used to advance partisan and particular interests rather than the public interest? By contrast, we might be more concerned about the po­tential tyranny of the bureaucracy—about government regulation and oversight becoming enmeshed in our daily lives, 30 These different visions also capture a tension that Michaels himself recognizes—between accountability and expertise. See id. at 553 n.169. Do structures that facilitate the former too often devolve into partisan hackery, or should we be more concerned about structures that facilitate the latter, empowering bureaucrats to complicate and stymie so­cial and economic exchanges? The debate between these two values is a constant in ad­ministrative law, and to say that each should be present in some measure gives us little guidance in resolving conflicts between the two, much less when confronting questions of institutional design. or being insufficiently attentive to popular preferences. The problem may be less one of con­centration than of scale and lack of accountability. 31 To refine the point further, we might seek to constrain law enforcement author­ity as much as possible, ensuring the use of force and powers of arrest and detention are governed by strict rules and demands of justification. But we might be comparatively less worried about regulatory power. If the latter were our concerns, robust centralized and political control would form the bulk of our answer.

And yet, in scholarly debates over the separation of powers, no sys­temic account emerges. Scholars tend to focus on how to recalibrate the dynamics of executive governance in light of discrete abuses or mistakes, or in service of particular regulatory agendas. The identification of an executive abuse either reflects the author’s particular preference for one of the types of government power (presidential power versus civil servant power, for example) or a conclusion that the exercise of a certain type of power has simply gone too far—a conclusion very much in line with Michaels’s call for balance. But these calls for recalibration have not pro­vided us with a comprehensive theory of abuse of power. 32 As Adam Cox and I have put it elsewhere, “If the last two decades of scholarship prove anything, it is that the appropriate level of centralization cannot be determined in the abstract . . . [but rather turns] on how the relevant institutions operate in practice.” Cox & Rodríguez, supra note 16, at 183. By centralization, we mean concentration of power and decisionmaking in agency heads and political leadership, rather than diffusion of power through the bureaucracy.

Michaels’s riposte to this indeterminacy could be that, because abuses of power can come in various forms, we need a system that nur­tures tensions among competing values. Because reasonable people can disagree about what sorts of government actions constitute abuse, a sys­tem that keeps all potential abusive tendencies in check will not satisfy everyone all of the time, but it will keep government action in bounds, generally and broadly understood.

This approach, and the recalibration tendencies of the literature de­scribed above, are certainly consistent with the claim that separation-of–powers problems must be resolved in pragmatic and context-specific ways—the approach I have defended elsewhere in work with Adam Cox. 33 Id. at 175 (explaining we cannot evaluate President Obama’s decision to defer re­moval of unauthorized immigrants as potential abuse of power without understanding his­tory and institutional context). But Michaels’s particular goal of maintaining rivalrous government sweeps too broadly. In any given struggle, the pursuit of tension will not provide a decision rule or principle to guide resolution of a controversy. In fact, keeping government power in check may sometimes require disa­bling one of the pillars of the administrative separation of powers—a pos­sibility highlighted below with an example from immigration law. On this view, the relative strength of the different types of power should wax and wane, rather than remain in some sort of rivalrous harmony. Preventing government abuses may in some instances be inconsistent with the exist­ence of rival powers.

We can take Michaels’s approach to the civil service as an example of the limitation of a theory based primarily on the preservation of rivalries. Though he insists that he is self-consciously presenting ideal types and acknowledges that each of the powers he defines might not function in the manner he suggests, 34 See Michaels, supra note 1, at 535, 546 (acknowledging civil servants are not nec­essarily always upstanding and capable). Michaels clearly has a soft spot for the civil ser­vice. One could read his account and conclude that he believes it should largely be insulated from control. His deep concern about the threat of privatization seems to stem in large part from the way public–private ar­rangements have simultaneously displaced the civil service and begun to change its character, lending his account of the civil service a romantic quality that may have produced an overvaluation of its strengths.

Michaels begins by styling civil servants as counterweights to political officials (the disinterested judiciary of the administrative separation of powers). 35 See id. at 541 (deeming civil service a potential “institutional rival”). On his account, civil servants, informed by a sense of profes­sional role or identity, will push back against the tendency to skirt laws and to promote hyperpartisan interests. 36 See id. (comparing relative freedom between civil servants under premodern Spoils System, in which government workers were often required to internalize political agenda of patrons, to modern civil servants and their ability to speak truth “without fear of serious reprisal”). Civil servants may not be di­rectly accountable to the public, as political officials are in theory, but that very insulation dampens the risks associated with accountability and thereby serves the overarching goal of constraint. 37 In this sense, law and politics exist in tension with one another, with the presence of one restraining the excesses of the other. See id. at 566 (suggesting civil serv­ants and fragmented administrative power may curb risk of overly broad Chevron deference that may represent “dangerous transfer of constitutional powers to the (unitary) Executive”). Civil servants can help insist that political leadership act fairly and according to law, rather than preferring their constituencies and advancing their own agendas without regard to the costs. The bureaucracy’s advantages along these lines stem not only from their insulation but also from their institutional memory and accumulation of knowledge and expertise over time. Whereas political officials strategize with short time horizons in mind, civil servants play the long game.

The problem with this stylized account is that a compelling counter­story can be told about the civil service. 38 For each of the threats he suggests privatization poses to the civil service, plausible counterarguments could be made. For his account of those threats, see id. at 580–83. Michaels argues, for example, that government power channeled through private conduits becomes harder for civil society to monitor. Id. But it could be that the disciplining effects of the mar­ket will produce better outcomes, at least in some instances, or that vigorous inves­tigative journalism will do as good a job of monitoring contractors as NGOs do of watching the government, or that private contractors can themselves be overseen by government ac­tors. Of course, this debate cannot really be resolved in the abstract and without a sense of what types of abuses are most important to prevent. What is more, most of his critique of con­tracting implicitly takes a side in the battle between power and constraint, concluding that constraining government is more important than effective and efficient government. For a discussion of this tension, see infra note 58 and accompanying text. Because power can serve the interests of the public as much as constraint and may well be best served by delegating power to the private sector, it becomes difficult to reject private contracting out of hand, even assuming it sidelines the civil service to a meaningful degree.

Michaels also cites contractors’ greater tendency to “embrace agency chiefs’ political priorities” as a reason to be suspicious of them. Id. at 582. But even assuming that civil serv­ants will not be inclined to advance their own ideological views in their work because of the absence of financial incentives for doing so, it is by no means obvious that we would not generally want the bureaucracy to follow the preferences of its politically accountable lead­ership, at least where those preferences do not amount to extreme rejections of evidence-based decisionmaking. First, civil servants’ insulation from politics relative to political appointees’ is arguably only a matter of degree. Civil servants transcend administrations and cannot be fired at will in the same way political appointees can be. But depending on reg­ulatory context, they may skew in one partisan direction or another, even if their political biases only manifest themselves subconsciously. 39 Career officials within the Environmental Protection Agency may be more likely to support the protective policies of Democrats, whereas line agents in law enforcement agencies may be more likely to support the robust enforcement policies of Republicans. More importantly, even if civil servants do not act on partisan interests in any direct sense, they are arguably more likely than political appointees to privilege the institutional interests of the executive branch. Their long-term professional interests and identities are linked to the stature of the agency or branch they inhabit. As such, they arguably are more likely to be partisans of executive power, whatever its manifestation, than political appointees whose primary interests will be in substantive policy out­comes. The civil servant’s inclination to favor regulation and bureaucracy presents a particular danger or problem if the abuses associated with con­centration of power consist of too much regulation or presence of gov­ernment in daily life. Civil servants almost certainly have more knowledge and understanding than the political appointees with whom they work, but they are probably more likely to become hidebound, un­creative, and rule-bound in the extreme than officials whose goals entail enactment of a particular agenda.

Of course, none of this critique of the civil service will be true all the time, and it may be a caricature to the same degree Michaels’s concep­tion amounts to romanticization. But the possibility of these competing narratives means that, in thinking about how to balance civil servants’ power with the authority of political appointees, it may be appropriate to disable the former to augment the power of the latter. Even if we make calculations of this sort only in context and not as a matter of general theory, we cannot avoid defining the abuse we most seek to prevent. To decide whether the civil service should or should not be disabled in a particular context, we must have a particular view about which systemic values should triumph in the end.

Take immigration enforcement as an example, where a battle be­tween the political appointees of the Obama Administration and elements of the civil service has recently exploded onto the front pages in the form of controversy over the President’s decision to defer the depor­tation of millions of unauthorized immigrants. 40 See, e.g., David Nakamura, Obama Acts on Immigration, Announcing Decision to Defer Deportations of 4 Million, Wash. Post (Nov. 20, 2014), https://www.washingtonpost.
com/politics/obama-acts-on-immigration-announcing-decision-to-defer-deportations-of-4-million/2014/11/20/9a5c3856-70f6-11e4-8808-afaa1e3a33ef_story.html [
XD29-8F66] (describing President Obama’s executive action to offer legal status to mil­lions of undocumented immigrants and field agents’ critique of it). For an account of these relief policies, see Cox & Rodríguez, supra note 16, at 137–41.
In a recent article about the President’s enforcement power, Adam Cox and I use these initiatives as a lens through which to demonstrate how constraining government power can sometimes require disabling the civil service, 41 See Cox & Rodríguez, supra note 16, at 185–92 (showing Obama deportation policy cab­ins civil servants’ discretion by design). but implicit in our definition of constraint is a conclusion about the sorts of abuses that merit attention.

In the same spirit as Michaels’s work, we contend that critics who claim that the President has usurped congressional power focus on the wrong separation of powers relationship. 42 Id. at 109 (calling for attention to power dynamics within executive branch, ra­ther than between Executive and Congress). Instead of implicating the clas­sical question of how power is distributed between the political branches, the President’s initiatives truly concern the structure of power within the executive branch. 43 See id. at 141–95 (advancing “two-principals” model of decisionmaking, which frames executive power as “function of the imperatives of the president’s obligations un­der the Take Care Clause”). On our account, political appointees within the Department of Homeland Security (DHS) and officials in the White House conceived, designed, and announced the relief initiatives as re­sponses to the difficulties of controlling the civil service in its enforce­ment of the law. 44 Id. at 185–92 (describing failure of guidance documents to steer enforcement de­cisions of line-level officials). In formulating these initiatives, DHS may well have con­sulted or involved elements of the civil service within the agency, even as they sought to curtail the authority of line-level ICE agents. A complete account of the development of these in­itiatives would probably reveal the complexity of the civil service itself, with differ­ences among them defined in large part by their professional responsibilities rather than the fact of their membership in the civil service. Before the initiatives were developed, guidance docu­ments that set out enforcement priorities for the civil service to follow appeared to have minimal effect on their actions. 45 Id. at 189–90 & n.246 (citing public accounts of enforcement actions and data con­cerning types of arrests under guidance regimes). In response, the President and leadership with DHS created two programs that invited unauthorized immigrants who met certain criteria to apply for relief from removal and work authorization—mechanisms, we argue, that cen­tralized discretion over who was targeted for deportation and cabined the discretion of line-level enforcement officials in service of the enforce­ment agenda of a Democratic President. 46 For a detailed discussion of these dynamics, see id. at 133–41.

Assuming we are correct about this characterization, it then becomes possible to see the move to constrain the civil service not as a power grab or a refusal to enforce the law, but as a means of disciplining government power by making its exercise transparent and accountable. This conclu­sion stands in tension with the one that arguably flows from Michaels’s approach—that immigration line officers’ resistance to the enforcement guidelines that preceded the Obama relief initiatives, as well as to the initiatives themselves, embodied rivalrous government. 47 For examples of civil-service resistance to Obama Administration initiatives, see id. at 170 n.187, 187 n.237, 189–90. On this view, political officials disrupted a tension that ought to have been allowed to persist.

So how do we know which outcome to prefer, as a matter of the sepa­ration of powers? Descriptively speaking, the Obama relief initiatives re­flect the power of political leadership to centralize control, at least incre­mentally over time. On Michaels’s view, perhaps their emergence high­lights that the political arm of the administrative separation of powers is the most in need of constraint precisely because of its potency. But in defending the relief initiatives, my coauthor and I come down on the side of breaking the rivalry because of a belief in the larger systemic value of accountable government, especially in a setting that involves the coer­cive powers of arrest and detention. These episodes may well reflect a decline in rivalrous government through the disabling of the civil service, but in context we believe them to serve the goals of separated powers more than the rivalrous status quo that preceded them. The President’s deci­sion to reorganize the bureaucracy reflected a complex amalgam of insti­tutional, political, and humanitarian concerns, and there has been public backlash against these efforts in Congress and through litigation that re­mains pending as of this writing. 48 For discussion of this litigation, see id. at 145–46 nn.119 & 120. But we reject the notion that any con­ception of properly balanced power ought to preclude political leader­ship from determining how best to organize the bureaucracy to serve rea­sonable ends leadership has identified as vital.

Of course, if the rise of privatization not only constrains the civil ser­vice in certain domains or discrete interactions but also has begun to de­grade it through and through, Michaels’s alarm seems warranted. But maintaining rivalrous government for its own sake will sometimes, if not often, be misguided. We must be open to the possibility that an element of the administrative separation of powers might need to be sidelined in service of a broader systemic goal (whatever we decide those goals should be). But this openness does not mean that we must give up on con­strained government. In Part III, I turn to describe an alternative account of constraint and relate it to the different theories of abuse defined above.

III. Complexity as Constraint

Though I share Michaels’s general preoccupation with constraining the arbitrary exercise of government power—few would reject that basic premise—I am skeptical that such constraint depends on assiduously maintaining the particular structural relationships Michaels defines as the components of the administrative separation of powers. As I ex­plained in Part II, balance in a given setting can actually undermine the goal of disciplining power. But even if, in the main, we should attempt to foster competing interests in government in order to keep power in check, we need not be preoccupied with maintaining balanced rivalries of the very specific sort Michaels delineates, even if we can tie them to an original constitutional design. Instead, executive power in today’s world—a form of power that has grown into a colossus if viewed at a cer­tain level of abstraction—comes from the brute fact of government’s complexity. In this Part, I focus on explaining how complexity constrains. But to the extent we should turn this description into a set of prescrip­tions, I also suggest we would be best off designing institutions and prac­tices within the executive branch to create overlapping functions and interests, 49 For an analysis of the functions of overlapping jurisdictions, see Jacob E. Gersen, Overlapping and Underlapping Jurisdictions in Administrative Law, 2006 Sup. Ct. Rev. 201. For scholarly accounts of how to design internal executive branch procedures to constrain power, see generally Katyal, supra note 16, and Metzger, supra note 16. as well as mutually reinforcing forms of supervision, none of which need track the rivalries Michaels describes.

The complexity thesis arises from the fact that any actor within the executive branch, from the President down through the bureaucracy, faces a dense web of intertwined interests that will make the exercise of power slow and challenging enough to prevent thoroughgoing abuses. Even if we put to one side the constraints exerted by Congress in its su­pervisory role, 50 For a discussion of these constraints, see Cox & Rodríguez, supra note 16, at 209–10. by public opinion, and by the nongovernmental institu­tions highlighted by Michaels and other scholars, 51 See Goldsmith, supra note 21, at 51–82 (citing media and NGOs as check on Executive, especially in making public aware of secret actions); Posner & Vermeule, supra note 13, at 113–53 (citing politics and public opinion as constraints). the executive branch itself is comprised of a range of actors and personnel that will work to­gether to keep power in check. This combination of interests might not be able to forestall every discrete instance of overweening presidential power or bureaucratic tyranny, but it serves as an effective bulwark against both, though to different degrees.

To be sure, apropos of my critique of Michaels in Part II, these claims about complexity do not mean that we should not worry about whether power in a given setting has become overly concentrated in po­litical leadership, the civil service, or the private sector. The immigration example I raised depends on identifying and combating precisely this form of overconcentration and thus disabling a rival. Nor is the complex­ity thesis wholly inconsistent with Michaels’s account. The existence of institutions that replicate the three types of power he identifies is itself a source of the complexity, because each type of power has particular core characteristics that lead to the prioritization of different approaches to policymaking, as well as to preferences for different policy outcomes. But the alternative understanding I present in the balance of this Essay is more fluid and pragmatic by virtue of eschewing the creation and tend­ing of categories.

Complexity as a source of constraint operates most clearly if we are thinking about how to prevent the concentration of power in the President and his political appointees. The executive branch may be more powerful today in relation to Congress than at any other point in our history, but the control that the President him or herself can exert over its vast operations is limited. An energized President can certainly champion particular initiatives or take control through his advisors of arms of the government to advance elements of his political platform. 52 For an account of this form of presidential power, see Elena Kagan, Presidential Administration, 114 Harv. L. Rev. 2245, 2251 (2001) (noting regulatory activity of agencies increasingly became extension of President’s own policy and political agenda during Clinton Administration and arguing “statutory delegation to an executive agency offi­cial . . . usu­ally should be read as allowing the President to assert directive authority . . . over the exer­cise of the delegated discretion”). This power may be especially potent in an era of polarized politics and a dysfunctional Congress whose capacity to check the executive branch through legislation or oversight seems to weaken by the day. But even a well-run White House cannot hope to touch but a small fraction of the administrative state.

Of course, as scholars have well documented, modern Presidents have sought to expand the reach of the White House through centralized review of agency action, under the auspices of the Office of Information and Regulatory Affairs (OIRA) and the Office of Management and Budget (OMB). 53 See, e.g., Cass R. Sunstein, Commentary, The Office of Information and Regulatory Affairs: Myths and Realities, 126 Harv. L. Rev. 1838, 1840 (2013) (detailing work of OIRA as information aggregator and interagency coordinator). The processes these White House components superin­tend certainly ensure oversight of the administrative state and may aug­ment the power of the White House relative to a world without central­ized review. But the processes also facilitate agencies’ constraints over one another by giving them a voice in the work of other parts of govern­ment with overlapping or complementary responsibilities. These central­ized review mechanisms ensure that agencies with “equities”—Washington speak for having a stake in others’ work—are made aware of activity in other parts of the administrative state and given the oppor­tunity to provide their own views. 54 See id. (“We shall see that while the President is ultimately in charge, . . . numer­ous agencies are also involved, and they may well be the driving forces in the process that is frequently misdescribed as ‘OIRA review.’”); id. at 1841 (“Part of OIRA’s defining mis­sion is to ensure that rulemaking agencies are able to receive the specialized infor­mation held by diverse people (usually career officials) within the executive branch.”); id. (“[M]ost of OIRA’s day-to-day work is usually spent not on costs and benefits, but on work­ing through interagency concerns, promoting receipt of public comments . . . ensuring discussion of al­ternatives, and promoting consideration of public comments.”). For a detailed account of the multiple dimensions of the interagency process and suggestions for reforms to im­prove its efficacy in coordinating government action, see Jody Freeman & Jim Rossi, Agency Coordination in Shared Regulatory Space, 125 Harv. L. Rev. 1131 (2012). Though the numerous interagency processes that operate within the executive branch may be styled as co­ordination mechanisms to improve the overall efficiency of the branch, they also institutionalize competing interests and provide vehicles for conflict resolution in a manner consistent with consensus-building rather than the freezing out of one set of views by a more powerful constituency.

The presence of various sorts of experts within the executive branch also ensures a level of deliberation and consultation that works against the concentration of power in political officials. If we think of the Executive not as a political branch of government but as the branch tasked with the responsibility of day-to-day governance, then we can be­gin to ap­preciate the difficulty of the work the branch must do. The prob­lems with which it must grapple demand expertise. Agency officials of all stripes therefore must consult with others outside their immediate orbits, creat­ing lines of conversation between political officials and bu­reaucrats who possess specialized knowledge both inside and outside their agencies, and even with sources outside government. A similar dy­namic obtains within the components of the White House itself and be­tween the White House and the executive branch as a whole. Im­portantly, the sort of expertise that will inform and temper political judg­ments can come from political appointees, such as the head of the Office of Science and Technology Policy or the Council of Economic Advisors, as well as from the civil ser­vice or outside interest groups. To the extent the imperative to get things done requires a turn to expertise, a multi­plicity of actors that blend the characteristics Michaels identifies in cate­gorical terms will shape the be­havior of political officials.

Another sort of expert who helps constrain political or partisan pro­jects is the lawyer—a pervasive figure across the executive branch who acts as a check on the exercise of government power. Criticisms of the government lawyer’s role are legion and range from aggravation and out­rage that they slow down national security decisionmaking in particular, 55 See Goldsmith, supra note 21, at 230–31 (describing former Secretary of Defense Donald Rumsfeld’s extreme dislike of lawyers’ challenges to national security policy and discussing potential costs of “lawfare”). to alarm at how easily even the most ostensibly disinterested of them can serve as enablers of the expansion of presidential power. 56 This critique reached a fever pitch in the wake of the revelation of the Office of Legal Counsel (OLC) memos issued in the immediate aftermath of September 11, 2001, conclud­ing that interrogation techniques widely regarded as torture did not in fact constitute tor­ture. For a systemic critique highlighting lawyers’ compromised positions with the execu­tive branch, see Ackerman, supra note 14. For a response to these claims based on a thor­ough institu­tional analysis of the incentives of lawyers in the White House Counsel’s Office, as well as OLC, see Trevor Morrison, Constitutional Alarmism, 124 Harv. L. Rev. 1688 (2011). An exploration of the ins-and-outs of this debate is well beyond the scope of this Essay, but an important take-away for my purposes is that Morrison effectively demon­strates how ex­ecu­tive branch lawyers, in the main, do restrain political officials, even as serious failures to do so remain possible and require a response. Id. We might also have a productive debate about the distribution of influence among lawyers within the executive branch. If we think of the OLC as the most “objective” or apolitical group of lawyers within the branch, we might worry that a decline in the Office’s influence over national security judg­ments will lead to less sound decisions by the President and the agencies. For an ac­count of OLC’s declining in­fluence during the Obama Administration, see Jack Goldsmith, The Decline of the OLC, Lawfare (Oct. 28, 2015, 6:11 pm), [https://per]. For a still different critique of lawyers within the execu­tive branch, see Margo Schlanger, Infiltrate the NSA, Democracy, Winter 2015, http://www.democracyjourn [] (ar­guing lawyers’ excessive focus on legality of agency action obscures deep­er questions of right and wrong and can result in undervalua­tion and weak defense of civil liberties).
But even though these criticisms ring true some of the time, the presence of law­yers, whose approval must be sought when ambiguities arise about the scope of an agency’s or official’s power, facilitates another form of delib­eration—about the legal legitimacy of state action. Whether those law­yers are polit­ical appointees or civil servants, they operate according to a conception of their professional role, which entails a general awareness that their re­sponsibilities include keeping policymakers within adminis­trative, statu­tory, and even constitutional bounds. These constraints oper­ate on all types of agency officials, but they are arguably most potent in restraining pursuit of partisan political objectives through the apparat­uses of the ad­ministrative state.

But while the role of complexity in constraining presidential and po­litical power seems clear, it seems less obvious that complexity constrains if our definition of abuse revolves around controlling the growth and tyranny of the bureaucracy. On the one hand, complexity can slow and even prevent government action, particularly to the extent structures within the executive branch facilitate oversight of one agency by others, either through White House coordination mechanisms or through over­lapping jurisdictions. By working against regulation, complexity might be said to diminish the role of government in our lives.

At the same time, a large government that cannot easily be con­trolled by political principals because of its many moving parts will in some instances engage in inefficient, or even oppressive, reg­ulation, ei­ther out of habit or a deep sense of institutional mission. On this view, complexity is itself the source of abuse. A related worry might arise if our goal was not to reduce the scale of government but rather to make gov­ernment better—to produce more regulation that more effec­tively ad­dresses social problems. Complexity could very well be the regu­lar and predictable enemy of this objective, 57 See Michaels, supra note 1, at 572–73 (discussing critics who bemoan that “admin­istrative separation of powers has gone too far” by preventing agencies from regulating). too, though this goal has less to do with identifying what constitutes abuse of power and more to do with how to maximize government’s performance.

This range of concerns reflects a dilemma that Michaels himself identifies as central in separation of powers doctrine and theory—strik­ing the proper balance between power and constraint. 58 Regulatory initiatives can help tinker around the edges of this dilemma by prompt­ing regular evaluation of whether an appropriate balance has been struck. For example, with Executive Order 13,563, the Obama Administration initiated a process of “look back” review, calling for agencies to “facilitate the periodic review of existing significant regula­tions” and to “consider how best to promote retrospective analysis of rules that may be outmoded, ineffective, insufficient, or excessively burdensome, and to modify, streamline, expand, or repeal them in accordance with what has been learned.” Exec. Order No. 13,563, 76 Fed. Reg. 3821 (Jan. 21, 2011). Do we want gov­ernment to be slow and complicated and therefore constrained, or do we want government to get things done effectively and therefore to be flexi­ble and nimble? Complexity may seem to fit better with the former ra­ther than the latter. But it actually has virtues regardless of what one be­lieves about the desirability of government regulation. If our goal is to con­strain the growth of the bureaucratic state because of its potentially tyran­nical behavior, then we would seek to attack complexity through po­litical control, but only if the political principals in office share an an­tiregula­tory agenda (otherwise they might use political control to expand the state’s power). And if our goal is to promote good and vigorous gov­ern­ance, then we similarly might prize political direction from the cen­ter, but again only to the extent the principals share our beliefs about the proper role of government. Complexity, then, might be the best way of hedg­ing our bets. A government that must integrate numerous ways of thinking and different sorts of institutional actors may work slowly and inefficiently, but it will do a reasonable job of mediating between widely divergent world views about the purposes of government while ensuring that no one ideology drowns out all others.

But even if this crosscutting account of complexity does not per­suade, it still seems clear that sustaining particular types of rivalries for rivalry’s sake will neither be necessary all of the time nor sufficient all of the time to constrain government abuse, regardless of how we define abuse. And even if we are generally disposed toward vigorous govern­ment, or checked and cabined government, tradeoffs between power and constraint (whether of political officials or bureaucrats) will be best ad­dressed with attention to institutional detail. Our commitment should be less to recreating the Madisonian vision of types of power than to nurtur­ing two important possibilities: the existence of multiple sites of deci­sionmaking and interlocking spheres of influence on the one hand and the presence of coordination mechanisms and principles for deci­sionmaking to break impasses among competitors on the other. How these should be initially designed and then adapted over time to changes in the regulated world will depend on context. Given that we all have general intuitions about government power, this move away from the sort of general theory Michaels ventures may be frustrating. But it offers a better frame of mind with which to consider the organization and utiliza­tion of an executive power vast in scope.


Though it may not have been his aim, in An Enduring, Evolving Separation of Powers, Michaels displays why the dramatic expansion of ex­ecutive power over the last century does not present an existential threat to our system of constitutional governance. Through institutional adapta­tion, the types of power that serve as rivals to one another remain live within the executive branch and therefore keep the Executive con­strained in some general sense. Michaels may be right to identify privat­ization as a severe threat to the viability of the civil service and therefore to these adaptations of the modern administrative state. But in any fight to maintain constraints on executive power, it matters less that we find analogues to the three types of power he identifies as sitting in judgment of one another, than it does that we design our institutions to contain a healthy mix of overlapping and competing equities and interests. Of course, even under this more fluid approach, difficult separation of pow­ers questions will arise when the system’s competitors reach an impasse that must be broken. And if we do not specify what precisely it is that constitutes abuse of government power—whether it is excessive political control, bureaucratic tyranny, or something else altogether—then we will not have the resources to mediate these disputes.

In the end, these debates over the structure of the executive branch may be a distraction from the more fundamental threat to our system of government—whether to the dynamics of power or those of constraint. Our supine Congress and the polarized politics that bedevil the country today seem more worthy of the label “constitutional crisis.” Addressing these developments is neither Michaels’s project, nor mine in this Essay, though a response to Michaels’s work could easily have argued that the degradation of the Madisonian Congress—the deliberative, popular as­sembly with the most affirmative power to address social problems—pre­sents the real separation of powers dilemma of our time. 59 In this sense, Michaels’s celebration of civil society as the analogue to Congress in the administrative separation of powers is misplaced. Civil society may play an important and public-oriented role in constraining executive power, but it is in no way a substitute for a popularly responsive legislative body. This limitation of civil society remains true whether we prize Congress’s role in supervising the Executive, or in solving the nation’s problems and keeping its government properly funded. On this view, and considering reasons I explore in Part III, supra, the separation of powers crisis is no longer, if it ever was, the breath and centrality of executive power. The real problem rests in the deteriora­tion of the Madisonian framework as a system for channeling popular sovereignty. Like many scholars who focus on the executive branch as a lawmaking institution, we write to improve a second-best world. 60 See, e.g., Cox & Rodríguez, supra note 16, at 171 (observing debates over how to struc­ture executive enforcement in immigration law and advance immigration policy through ad­ministrative action constitute second-best strategies as compared to world with Congress will­ing and able to legislate in cooperation with executive branch). But even if this were our ani­mating concern, we would still be wise to follow Michaels’s general pur­suit: seeking adaptations that will ensure our system remains consistent with very basic notions of how a government predicated on popular sov­ereignty and the rule of law maintains its legitimacy.