In the first annual review of the Obama presidency, Pete Rouse, a senior White House adviser, identified the difficulties that the president had encountered in gaining control over his own administration. "[W]hen the economic team does not like a decision by the President," Rouse observed, "they have on occasion worked to re-litigate the overall policy." He then singled out Larry Summers, the director of the National Economic Council, as a major source of internal resistance, noting "deep dissatisfaction within the economic team with what is perceived to be Larry's imperious and heavy-handed direction." Going further, Rouse observed, "[O]nce a decision is made, implementation by the Department of the Treasury has at times been slow and uneven." All of these factors, he concluded, "adversely affect execution of the policy process" (Suskind 2011, 378).
Rouse's memo captures a fundamental and recurring problem that frames the central argument of this article: structural advantages said to follow from the constitutional design of the unitary executive--a first-mover advantage, a collective-action advantage, and an informational advantage--are of limited value without presidential control over administration, and, as even a brief historical analysis will show, presidential control over administration varies widely from one incumbent to the next. More specifically, I demonstrate that when the president relies on the established collective-action and informational arrangement, it often obstructs or compromises his first-mover advantage. Conversely, his first-mover advantage in a policy domain is strengthened when he restructures the established collective-action and informational arrangement.
Incorporating both a critique of several leading theoretical concepts in the field of presidential studies as well as an empirical analysis that ranges widely over the relevant history, this article proceeds as follows. In the next section, I examine the assumptions of the unitary executive framework. In the following two sections, I put the unitary executive framework to a test using cases from the early nineteenth century and the twenty-first century, respectively. And in the last section, I briefly discuss an alternative approach to studying the presidency. Throughout I reference the administration of public finance, broadly defined to include expenditures, taxation, and money and banking, or, in modern terms, fiscal policy and monetary policy.
The Unitary Executive Framework
Research on the presidency was long criticized for being too descriptive, atheoretical, and poorly designed. In contrast to other areas of political inquiry, systematic approaches, scholars claimed, were absent in presidential studies (Wayne 1983; King and Ragsdale 1988; Edwards et al. 1993). Accordingly, Richard Neustadt's book, Presidential Power, the leading work in the field for several decades--and one that emphasizes the complexities of the modern president's leadership position, the importance of bargaining, and the role of individual skill in maximizing presidential power within a governmental structure of "separated institutions sharing powers"--came under attack as the main target of this critique (1960, 39). The problem, Terry Moe argues, is that "Neustadt did not offer a coherent, well-developed theory, but rather a loose set of ideas ... that were not formulated with much precision" (2009, 703).
Taking direct aim at the perceived shortcomings that Neustadt's work represents, the so-called rational choice "revolution" in presidential studies has proceeded, in Moe's words, to bring "simplicity, clarity, logical rigor, and deductive power" to the study of the presidency (2009, 704). Although each contribution stands on its own merits, this line of research as a whole elaborates a set of structural claims that allegedly follow from executive unity. The Constitution, these scholars contend, endows the president with a first-mover advantage, a collective-action advantage, and an informational advantage relative to the other branches of government. Kenneth Mayer, for example, emphasizes that the president can take "advantage of his ability to move first" (2001, 152). (1) David Lewis and Terry Moe stress, "Presidents are not hobbled by collective action problems. Supreme within their own institutions, they can simply make authoritative decisions about what to do and then do it" (2014, 383). (2) And Brandice Canes-Wrone highlights the president's ability to use an "informational advantage to try to convince the Congress and electorate that his preferred policy is the one that will advance their interests" (2006, 32). (3)
With these structural advantages in hand, this literature cuts through much of the complexity that prior research had dwelled upon. The executive branch becomes a unitary actor--the president--operating in relation to Congress or the Court. The policymaking process is modeled as an iterated game, starting over each year or congressional term. And the parameters of the policy realm of interest, its administrative operations, and presidential control all remain fixed, impervious to change from one period to the next. From this perspective, the president simply employs one fixed structural advantage or another to shift the policy status quo and variation is then accounted for with reference to peripheral institutional conditions such as party control of the presidency, congressional support, and public approval.
William Howell's work, Power Without Persuasion, is a leading example of this approach, as it emphasizes how all three structural advantages enhance presidential power via executive orders. "The most important is that the president moves policy first and thereby places upon Congress and the courts the burden of revising a new political landscape," Howell reasons. Second, "the president acts alone," unconstrained by Congress' collective-action problems. And third, "the president often knows more about policy matters than do members of Congress, and he uses this fact to his advantage when deciding to act unilaterally." These structural advantages, Howell concludes, help to "[distinguish] unilateral powers from all other sources of influence. In this sense, Neustadt is turned upside-down, for unilateral action is the virtual antithesis of bargaining and persuading. Here, presidents just act" (2003, 14-15, 101-02).
Although Howell's work is an important contribution to our understanding of how the total number of executive orders issued varies in relation to certain institutional conditions, the unilateral action framework is analytically narrow and temporally restrictive. To begin with, consider the assumption underlying the first-mover advantage. The president acts first due to the design of his office, Howell reasons, "and thereby places upon Congress and the courts the burden of revising a new political landscape." Taking this reasoning further, it becomes impossible for incumbents to constrain their successors, as the assumption is that all presidents exercise the same first-mover advantage with the policy-making process starting anew each congressional period. Remove this structural conceit, however, and the first-mover advantage narrows considerably. The notion that all presidents are equally equipped to make and remake policy, reversing their predecessors' actions as they please, overlooks the process by which shifts in policy and new administrative arrangements are institutionalized. In other words, presidential initiatives may constrain Congress or the Court, but presidents who create a "new political landscape" also constrain their successors. (4)
Next, consider the two assumptions that "the president acts alone" and "often knows more about policy matters than do members of Congress." Here, Howell grants the president the benefit of a collective-action advantage without addressing the cost of information. Yet the simple fact of the matter is that if the president acts alone he is unlikely to have an informational advantage. Conversely, if he has an informational advantage there is likely a collective-action cost. Howell sidesteps this issue by ignoring the absolute cost and assuming that whatever it is it must be less for the president than for Congress or the Court. Or, to put it differently, he assumes that the president maintains structural advantages relative to the other branches of government. However, like the first-mover advantage, the benefit of a relative inter-branch collective-action or informational advantage becomes more complicated if we introduce realistic assumptions about "bounded rationality" and "durable shifts" in governance and policy (see Simon 1957; Orren and Skowronek 2004). The president's control over his own administration, for example, is always necessarily incomplete due to limited time, limited knowledge, and a constitutional system comprising multiple structural constraints that make it difficult to change existing institutional arrangements. (5) Accordingly, institutions within the executive branch may provide the president with an informational advantage relative to Congress or the Court, or they may use that information to constrain an incumbent with preferences different from their own. At the same time, as new institutions are created or as older ones are restructured, collective-action costs change. In certain cases, even if the president maintains a relative inter-branch advantage, the absolute cost may be so high that the entire government fails to act, leaving the policy status quo in place or allowing it to "drift" (see Hacker 2004; Hacker and Pierson 2014).
To appreciate the full extent of the problems with the unitary executive framework in operation, consider, for example, the president's structural advantages during the Hayes administration, a fair test if these advantages derive directly from the constitutional design of the office. Rutherford Hayes assumed...