The Dual Executive and Unilateral Power
Brandon Rottinghaus
Associate Professor
Department of Political Science
University of Houston
Abstract
The study of presidential unilateral powers profiles a powerful president and charts the shifting
balance of powers between the executive and legislative branches. While presidents do rely on their
constitutional powers when issuing unilateral orders, they also justify their actions based on
Congressional statute or other Congressional sources of authority. The puzzle, then, is whether
unilateral action can be a form of power sharing with Congress rather than acting alone. Given that
the Constitution requires the two branches to share in policy making power, when taking unilateral
action are presidents more likely to act in accord with or against, the will of Congress? Our view is
that the unilateral powers literature lacks an explanation of when and how presidents use their
unilateral orders as part of a system of both shared and separated powers. Presidents behave
differently depending on their political environment and presidents issue different types of unilateral
orders in different circumstances in reaction to specific political conditions or events ― sometimes
cooperatively, sometimes contentiously. In this abbreviated paper (from our book), we examine two
distinct phases of the policy process (before Congress acts and after laws are passed) to determine
when and under what conditions presidents are more likely to act alone. We find that the president
is more likely to issue command and adaptive implementation orders when relying on broad
substantive discretion and constitutional authority, when party polarization is higher, or when the
distance between the presidents and the Senate filibuster pivot is larger.
Unilateral power, like executive power, is diverse. As Mansfield (1989) argued, the word
executive has two meanings. The executive serves to carry out the will of the legislature and to enforce
that will and therefore must exercise its own discretion. Scigliano (1989) notes that scholars have
“embraced two conceptions of executive power: one makes the president subordinate to Congress,
whereas the other allows him to be autonomous and self-directing within broad limits” (238).
Justice Marshall, in Marbury v. Madison (1803) wrote that, “By the Constitution of the United States,
the President is invested with certain important political powers, in the exercise of which he is to use
his own discretion” but as Marshall and later courts have recognized, “a ministerial act which the law
enjoins” must also be carried out by the president. The same is true with regard to unilateral power.
1
We cannot assume a singular nature to the use of presidential unilateral action. The president can
use a unilateral order to pursue his administration’s policy interests without Congressional support
or exercise his duties and obligations related to the management of government equally with
Congress.
A New Approach to Unilateral Power
In this paper (and the larger book project) we propose a new approach to the study of
unilateral powers by providing a framework for studying unilateral orders within the executiveCongressional relationship. We argue there are dual roles for the executive: an “independent”
president who pursues his agenda and a president who “administers” government functions often
with the support of Congress. The study of the exercise of unilateral power has not considered fully
how these dual responsibilities jointly shape presidential unilateral action. The use of unilateral
orders is dynamic, where the president functions both as adversary or ally of Congress and as an
independent agent or as routine manager. Therefore, the next step in understanding the use of
unilateral power is to investigate the “shared” nature of these orders through the dual functions of
executive authority and power as framed by the Constitution. The type of order a president issues,
the political circumstances surrounding it and the specific executive function he is operating under
will largely dictate when and under what circumstances an order is issued. Theoretically, then, we
should see differences in the findings based upon the nature of the president’s authority to act, the
timing of the order in the policymaking process and the intent of the president’s action.
This conceptualization tells us something important about how presidents make policy,
especially unilateral policy, in a shared power context. The nature and timing of unilateral orders in
the political system are not well understood since these orders are often treated as monolithic. In
fact, these orders are used for different purposes at distinct moments in the policy process. The
substance, tone and timing of the orders identified in this book can clarify the use (and abuse) of
2
presidential executive power. Indeed, scholars presume that unilateral power allows the president to
act without restraint and when the need suits him. Although scholarship has demonstrated that the
president acts with possible Congressional or judicial reactive restraint in mind, fears of an
aggressive White House marauding over their constitutionally assigned window of responsibility
remain constant. This helps to put some of these fears to rest by identifying the moments of shared
political power between the branches that both allow presidents to act with dispatch, as the Framers
intended, and to work with Congress as a means to navigate and implement public policy. While
motivation, opportunities and execution of overreaching executive unilateral actions has occurred,
and will no doubt continue, this book places these moments in the proper context of the shared
power arrangements outlined by the Constitution.
Presidential Power in the Unilateral Context
Most studies of presidential power rightfully genuflect on the role that Richard Neustadt’s
Presidential Power (1990) played in reshaping the discipline’s views of how presidents exercise power.
Strictly unilateral uses of power are described by Neustadt as a “painful last resort” to be used
expediently only after bargaining has failed (11, 24). Challenging Neustadt’s conceptualization of
presidential power, several scholars confront Neustadt’s view of a “weak” president arguing it is at
odds with the formal powers presidents exercise when engaging in unilateral action (Howell 2003;
Mayer 2009; Waterman 2009). It is clear that unilateral orders are powerful (and often dramatic)
examples of the president’s ability to command (Cooper 1986; Shull 1993; Mayer 2001). Cooper
(2002) argues that “there is virtually no significant policy area in which presidents operate that has
not been shaped to one degree or another by the use or abuse of these tools” (ix). Presidents have
the capacity to establish rules, norms and policy consistent with their own preferences and without
input from Congress. In several instances, including institutional executive power, regulatory
authority, budget authority and civil rights, “the president’s ability to move first, combined with
3
Congress’s relative inability to respond effectively, tilted competition in favor of the executive”
(Mayer 2001, 112). Unilateral orders allow the president to “do what Congress could not or would
not do” (Neighbors 1963, 118). According to Fleishman and Anfses (1976), executive lawmaking
by unilateral order is powerful enough to undermine “democratic decision procedures, and threatens
the rule of law” (40; Hebe 1972).
The suggestion that unilateral orders are powerful policy tools that can be used at the
president’s discretion gave rise to the idea that president may use these powers to short circuit the
separation of powers. By formal arrangement, the growth of institutional power, and the ambiguity
of the Constitution, “presidents can and do make new law ― and thus shift the existing status quo ―
without the explicit consent of Congress” (Moe and Howell 1999, 133). Unilateral orders allow
presidents to act alone in an “efficient and alternative manner” compared to the legislative process
(Krause and Cohen 1997,462; Deering and Maltzman 1999). Described by scholars as the “strategic
model” (or “evasion hypothesis,” Martin 1999), presidents are strategic actors who issue unilateral
orders to bypass or evade Congress (Morgan 1970; Nathan 1983; Peterson 1990; Fine and Warber
2012). The growth of unilateral action chiefly during emergencies (wars or economic crises) is a
direct outgrowth of the deference granted to presidents during these times by Congress (Cooper
1986, 236; Howell 2005) and the opportunities they present for presidents to act alone (Howell and
Pevehouse 2005). These temporal entreaties beget further, seemingly innocuous, expansions into a
wide range of policy making, in both foreign and domestic policy (Moe and Howell 1999).
“Shared” Powers in the Unilateral Context
The emphasis on the president’s formal powers guides our contemporary estimations about
presidential power but may underestimate the degree to which “unilateral” powers are shared
between the branches with the president and Congress working together. This notion of shared
powers is true even in the unilateral power context. The issuance of a unilateral order itself does not
4
mean that the president is necessarily fulfilling his own prerogative. In reality the president may be
executing the will of Congress through prearranged policy or acting administratively to continue
government functions with the tacit consent of Congress or in the context of Congressional
disinterest. For instance, Shull (1997) suggests that the issuance of unilateral orders is “influenced
by interactions with Congress” (97) and presidents “seldom act without input and constraint from
others” (98). Shull (1997, 123) also previews this line of argument when he suggests that presidents
often issue executive orders that are “highly supported by Congress, they [presidents] appear to be
using them as a way of implementing legislation with which they agree or for routine matters rather
than seeking alternative policy adoption.” Dickinson (2008) argues that the phrase “unilateral
action” implies presidents work “alone” but this belies the fact that that presidents must bargain
with policy stakeholders inside and outside of the executive branch widely even when issuing
unilateral orders. Two of the three examples in the beginning of this chapter show an extended
working relationship with Congress behind the façade of the unilateral action.
If the president uses unilateral orders to carry out dual executive functions he should be
alternatively acting independently and as an administrative agent in the legislative process. For
instance, scholars have suggested that presidents may preempt legislation they dislike with a more
moderate unilateral order (Howell 2003), where they are more likely to do so when Congress is
weaker or when they are unable to pass legislation through Congress (Fleishman and Anfses 1976).
Yet, the president may agree with the proposed legislation and help Congress to overcome possible
collective action problems. Presidents may also use their orders to implement legislation in specific
ways after it has been passed, where the need to bargain with Congress is lessened. Acting
unilaterally, the president may rely on statutes or their own authority to execute their current political
agenda or modify a prior order. Because legislation may be complex, incomplete or vague upon
passage, presidents may issue interpretive orders to remedy these problems, both with and without
5
Congressional support. More frequently, presidents use unilateral orders for carrying out routine
bureaucratic functions (Schramm 1981), with nearly 60% of executive orders are used for routine
purposes (Warber 2006, 39).
Underpinning the theory that presidents employ different types of unilateral orders to
execute their dual executive functions is the notion that the political environment dictates the use of
these orders. At their heart, theories of unilateral action often draw the distinction between
“unilateral” and “bargaining” actions, which are sometimes difficult to show because the president
acts under both conditions. The president’s unilateral action is also subject to prior policy restraint
since Congress initially limits the policy action that presidents can undertake (Shull 1997, 98). The
president may direct the action relying on his own authority to effectuate policy change or
implement policy. The president both commands using his inherent authority and acts as
administrator by carrying out delegated and necessary functions. This duality in executive function
suggests that there is a need to examine unilateral activity with respect to the roles that the executive
plays within the larger institutional dynamic between the two branches of government (see Marshall
2011).
Rethinking Unilateral Power
The continuing debate about the role of unilateral orders provides an opportunity to rethink
our approach to unilateral powers. Indeed, Mayer (2009) suggests that “the disparate and conflicting
results [in prior studies] strongly suggest that we do not have a complete understanding of the
interbranch dynamics of unilateral actions” (441). We argue a “shared power” arrangement is not
anathema to presidential unilateral powers. Recent work on the unilateral presidency finds that
presidents can (and often do) act unilaterally with respect to their political position but that the use
of orders is limited by the president’s shared structural, institutional constraints. For instance, Mayer
(1999) argues that certain orders are “hybrid” since they “constitute a potential reservoir of
6
independent authority, but one that presidents do not use without regard to circumstance or
consequence” (448). Likewise, Howell (2003) argues that “institutional constraints lie at the heart of
a theory of direct presidential action for they determine what presidents can actually accomplish.
The freedom the president has, at any moment, to unilaterally set public policy depends critically
upon how the other branches of government will respond” (176). Moe and Howell (1999) argue
that presidents are “political animals” and are keenly aware that extreme action taken in one policy
domain may jeopardize their effectiveness in another. Thus, presidents may self-check, moderate
their actions and “take much smaller steps than their de facto powers would allow” (153).
A New Approach to Unilateral Action
We propose a new approach to the study of unilateral powers by analyzing its application
within the separate and shared powers framework. This framework allows us to consider how the
use of unilateral orders tracks the dual role of the executive as “independent” (where presidents use
unilateral orders to advance their own policy objectives) and “administrator” (where the president
keeps the wheels of government oiled or acts as the agent of Congress). The “independent”
executive fulfills the expectation in existing scholarly research which shows presidents acting against
Congress and on behalf of his own agenda. “Administrator” presidents use their unilateral authority
as expressly connected to Congressional intent or as a means to carry out routine executive
functions. This work helps to delineate scholarship which both shows that much unilateral action is
of a routine nature and scholarship that demonstrates that presidents are aggressive agents in acting
unilaterally. To highlight these differences, this book explores the decision-making process
pertaining to unilateral orders within three separate and shared policy creating arrangements
focusing on the circumstances under which presidents issue unilateral orders and the key factors that
influence the decision to build Congressional preferences into unilateral action.
7
Our book develops an organizational model to chart the process of when and how a
unilateral order is used and, from this model, we derive several theoretical explanations for when the
president should act as “independent” or “administrator.” We create the context for analyzing these
roles based on whether the president’s institutional role is independent or shared with Congress. In
order to test our expectations concerning the dual executive, we consider the influence of multiple
integrated power arrangements including: the source of authority presidents use in justifying their
orders (Congressional statute, Congressional resolution, the Constitution or inherent executive
authority), the amount of delegation Congress gave the president with respect to each unilateral
order issued, whether the president used an executive order to preempt or facilitate Congressional
legislation on the same topic and how presidents use unilateral orders after legislation is passed to
“faithfully” implement law (or not). This approach demonstrates the dual role of the executive and
the connection to unilateral orders and fits within the system that requires shared powers via the
Constitution.
The Executive’s Dual Roles
No one would suggest that presidents are only either “clerks” or “commanders” (in
Neustadtian typology) ― rather presidents have evoked both types of interpretations of their power.
With an “independent” president, unilateral orders are used to put his own issues on the agenda and
to initiate policy goals consistent with an “energetic” executive (Rossiter 1956). This is the
expectation of scholars who suggest that presidents use executive orders to set a policy agenda, act
when Congress won’t, and preempt an action Congress is considering of which the White House
disapproves (Neighbors 1963). An “administrator” president uses orders as a means to manage the
affairs of government, about which the president is either agnostic with respect to the issue or shares
the views of his lawmaking partners. In this capacity, the president fulfills a central but largely
benign executive function of administering governmental maintenance by carrying out laws that
8
have been passed or initiating the policy implementation process. In the discussion below, we
functionally separate unilateral orders into these two categories to mirror the roles that presidents
play in the exercise of power.
FIGURE 1.1
Scale of Presidential Unilateral Action
___________________________________________________________________________
NSC Name Change
Less Unilateral
“Administrator”
Contractor Overtime Pay Order
More Unilateral
“Independent”
___________________________________________________________________________
The scope of presidential unilateral action, then, can be placed on a scale of action ranging
from more to less unilateral. Figure 1.1 illustrates this span on a simple scale and a discussion of
these two examples illustrate the differences. In the same week in February of 2014, President
Obama issued two executive orders of very different ilk. In one order, the President issued an order
which changed the name of the “National Security Staff” to the “National Security Council Staff.”
The order specifically notes that it “shall not be construed to impair or otherwise affect (i) the
authority granted by law to an executive department, agency, or the head thereof; or (ii) the
functions of the Director of the Office of Management and Budget relating to budgetary,
administrative, or legislative proposals.”1 Although issued by the “stroke of a pen,” the order raised
no concern in Congress and was little more than cosmetic in nature and as without objection from
Congress who gave him this authority to act. In another order (Executive Order 13658), the
President, as part of his promise to sidestep Congress and reduce income inequality, ordered the
Labor Department to revise overtime pay rules to “make millions more workers eligible for extra
1
Executive Order 13657, “Changing the Name of the National Security Staff to the National
Security Council Staff.” February 10, 2014.
9
pay when they work more than 40 hours a week” (Joachim 2014).2 Getting little traction in
Congress on legislation on raising wages, the President acted without Congress using an order in a
more unilateral way. The order was issued before Congress acted on the matter and raised the ire of
both Congressional opponents and the business community.
This range helps us to identify the scope and scale of presidential unilateral action. We
describe the two unique types here. As suggested, presidential “independent” unilateral orders are
used to further the president’s policy objectives or protect the power of the executive branch. For
these orders, we include: “command” orders (from issues on their political agenda) when presidents
act before Congress acts, orders which “preempt” Congressional legislation and orders that “adapt”
rather than implement law (where presidents issue an order to implement policy in a flexible way).
On the other hand, we identify three companion orders (issued at matched points in time, see Figure
2.2 below) where we suggest that presidents should be more likely to work with Congress to further
the policy process and engage in mutual and expedient movement of legislation into law – we label
these as the “administrator” orders. These include: “routine” orders (managing the affairs of
government), “support” orders (issued before Congress legislates but in concert with Congressional
preferences) and “implementation” orders (which faithfully enact legislation passed by Congress into
law). The scope of the president’s authority, the president’s political environment and the
institutional dynamic with Congress will dictate which of the types of orders the president issues at
each stage.
Consistent with “independent” executive orders, presidents should be more likely to issue
these when they are less able to get agreement on a policy direction from Congress or where they
disagree with Congress. For instance, when government is divided, an opposition Congress is less
likely to support the president’s initiatives and he may have greater incentive in these instances to
2
Executive Order 13658, “Establishing a Minimum Wage for Contractors.” February 12, 2014.
10
pursue an independent and unilateral policy (Deering and Maltzman 1999). Although presidents are
limited by an anticipatory response from Congress they are nonetheless willing to maximize power
and act against a recalcitrant Congress to achieve their policy objectives (Moe 1993). Presidents
often use unilateral orders to accomplish policy goals that Congress could or would not do
(Neighbors 1964). If unilateral orders are a way to make policy without Congress, we should see
presidents using them more frequently when there is greater interbranch policy disagreement. In
these instances presidents are less able to bargain (and legislate) with a divided and recalcitrant
Congress. Presidents running for reelection or in the second half of their term or second term are
often busily trying to convince Congress to work with them despite the fact that they are often
unwilling to do so.
Alternatively, using “administrator” orders, the president is more likely to work with
Congress when they have mutually agreeable objectives, when there is more support in Congress for
the action he is likely to take, or when Congress is disinterested in the outcomes and relies on the
expertise of the executive to act. Under these conditions the president and Congress fortuitously
share in policy development and can more easily achieve a satisfactory policy arrangement. For
instance, presidents may treat unilateral orders as “complimentary” to the legislative process (see
Gleiber and Shull 1992; Krause and Cohen 1997). If Congress and the president are closer aligned
either in terms of party or ideology (or both), it stands to reason that they would, in general, agree
more with each other and the president can use unilateral orders to their mutual benefit. Scholars
have discovered that when the president has greater party support, he issues more orders because he
is more likely to be in greater agreement with Congress (Shull 1997, 103; Gomez and Shull 1995).
Even when the branches are in potential conflict, the institutional responsibility to act should
transcend ideological disagreements for these orders in order to keep government running.
11
The Stages of Presidential Action
In the previous section we discussed how we discern the dual roles of the executive from the
Constitution where we see the branches sharing powers. In this section we map three stages that
chart our conceptualization of the timing of the president’s use of unilateral orders. As shown in
Figure 2.2, in the first stage (upper panel) (“Direct Action”) unilateral orders are the output of the
president’s primary action, where the president’s unilateral action occurs before the legislative
process begins. Congress may be in the policy development stage and direct action represents an
opportunity for the president to act first on information obtained by the executive branch.
Presidents may also have a specific policy objective in mind and do not want to wait for the
legislative process to initiate policy. In the second stage (“Legislative Process”), the president’s
unilateral order is issued during the legislative process related to bills that have already reached the
Congressional agenda. The president now has an opportunity to work with Congress on proposed
legislation where he may or may not find success or issue an order altering the trajectory of the
legislative process. In the final stage (“Final Law”), legislation has been signed into law and
authority delegated to the president which enables him to issue unilateral orders to implement final
legislation.
12
FIGURE 1.2
Model of Unilateral Orders and the Policy Process
The following sections expand on the nature of each of these orders with respect to the
president’s dual executive roles and provide examples to illustrate each specific order. Because we
employ distinct legislative or operational cut points, we argue all orders can be classified into one of
these time periods. We focus here on the first and third stages.
Direct Presidential Action. When the president issues a unilateral order in the first stage, he is
exercising action before Congress, generally at the policy incubation stage (Rudalvidge 2011). We
identify , the two types of orders the president issues at this stage as “routine” and “command” .
First, “routine” orders are primarily administrative or managerial in origin. This is an example of the
administrative presidency working to maintain the function of government and “keep the trains”
running. In our framework, these issues are not on the president’s agenda and are used benignly to
13
jointly manage government.3 For example, routine orders are used to make internal organizational
changes within the executive branch such as altering a pay scale or job description, make necessary
changes within the executive branch in order to comply with a judicial opinion, delegate authority or
reorganize agency responsibilities or reporting mechanisms. These are not likely orders that would
raise the ire of political opponents or Constitutional purists. The Obama Administration order to
change the name of the National Security Council Staff referenced above is a clear example of this.
Routine orders are primarily administrative. The purpose is generally, as Warber describes,
(2006, 142) “to accomplish a simple, noncontroversial, administrative task in the executive branch.”
For example, presidents use executive orders to create an investigatory board to solve labor disputes
between union-represented employees and their employers pursuant to the Railway Labor Act (45
U.S.C. 151 et seq; see also Mayer 2001, 76).4 Similarly, after the terrorist attacks of September 11,
2001, President Bush issued several executive orders providing for an order of succession within
each executive department in response to this national emergency.5 These orders provided for
agency continuity in the event of a change in leadership and, in total, sought to preserve the
continuity of government. Although the response to the terrorist attacks was foremost on President
Bush’s agenda, the change in succession wasn’t specifically discussed as part of his public agenda nor
was the order a request of Congress.
Unilateral orders which “command” are distinguishable from those used to administer
routine functions because they stem directly from the president’s public agenda and are used to fashion his
3
To be considered on the president’s agenda, the issue had to have been spoken by the president at
least three times in public. This constitutes a high bar for an issue to be on the president’s agenda
because the president has publically prioritized the issue. We utilized the Public Papers of the President
to document the president’s agenda.
4
The president’s authority is discretionary under Section 9A and mandatory under Section 10 of the
Railway Labor Act.
5
Executive Orders 13241-13247 were issued on December 18, 2001 and Executive Orders 1325013251 were issued on December 28, 2001.
14
own policy. More specifically, command orders signal to the Congress or the public the president’s
commitment to a larger initiative and, in some instances, set the agenda for Congress to follow. For
example, relying on Constitutional authority and his role as Commander-in-Chief, President Carter
issued Proclamation 4483 and Executive Order 11967 on January 21, 1977 pardoning Vietnam War
draft evaders one day after taking the oath of office. President Carter fulfilled his campaign promise
by acting quickly to execute the pardon but his unilateral action did not sit well with members of
Congress who had been working on developing a legislative approach during the prior session.
Command orders may also be used to centralize policy in the executive branch by establishing a
formal mechanism such as a committee or commission. In the day before President Carter’s final
day in office he issued Executive Order 12285 establishing the President’s Commission on Hostage
Compensation. If a presidential commission is of particular importance, it may be continued from
one administration to the next (which President Reagan did under subsequent executive orders).6
The president may also do more than continue the commission by using unilateral orders as a way to
amend, extend, and shape the composition of committees to allow the White House greater control
over the policy development process.
Executing Law. While Congress has the powers to make all laws, the president has the
constitutional authority to execute and enforce them. The statutory authority specified in the
language in the law that enables him to “faithfully execute” the will of Congress. This is the final
phase of the policy process and the one in which the president assumes Neustadt’s “clerkship”
because “no one else’s services suffice” (1990, 7). When unilateral orders “faithfully” implement
legislation, the president is acting in concert with Congress, following the direction specified in the
6
Executive Order 12307, “President's Commission on Hostage Compensation” authorized the
commission to conduct public hearings and gave the commission additional time to prepare the
report (46 FR 30483; June 9, 1981) and Executive Order 12317, “President’s Commission on
Hostage Compensation” gave the commission additional time to prepare the report (46 FR 42241;
August 20, 1981).
15
statute. To be considered “faithful” implementation of a law, the president must either (1) cite the
specific statutory authority of legislation passed in the Congressional session in which the order was
issued or (2) attribute his action in the statement or remarks accompanying the signing of unilateral
order (either a speech or a signing statement). Orders that implement policy action can involve
changing or enacting agency regulations, triggering a particular statutory action and responding to
emergencies (Gleiber and Shull 1992; Shull 1997, 2006; Cooper 2002; Kerwin 2003). Sometimes
presidents may use executive action to evoke multiple conditions as President Carter did by issuing
an executive order to trigger a particular statutory action and respond to a natural gas emergency
pursuant to the Emergency Natural Gas Act of 1977 which he had signed into law the same day.7
Presidents are generally expected by Congress to “faithfully” implement legislation involving
the management of policy within the executive branch such as the establishment of a commission,
an organizational or administrative change, delegation or assignment of authority. The authority to
implement a law is generally delegated by Congress to a division or agency within the executive
branch or to the president. 8 The president may perform his statutory duty by sub-delegating his
authority, as President Carter did to implement his obligations under the Diplomatic Relations Act.9
However, Congress could have (but did not) direct the president only to designate an individual to
carry out the duties under the law as in the Uniformed and Overseas Citizens Absentee Voting
Act.10
7
Executive Order 11969, “Administration of the Emergency Natural Gas Act of 1977”
(February 4, 1977; 42 FR 6791). Emergency Natural Gas Act of 1977 (Public Law 95-2, 91 Stat. 4).
8
For a debate on whether on the unitary executive and whether the president has discretionary
authority over executive branch agencies receiving delegated statutory authority directly see Lessig
and Sunstein (1994) and Calabresi and Prakash (1994).
9
Executive Order 12101, Privileges, immunities and liability insurance for diplomatic
missions and personnel (Nov. 17, 1978; 43 FR 54195). HR 7819, P.L. 95-393 Diplomatic
Relations Act Public Law 95-393, 92 Stat. 808; 22 U.S.C. 254a et seq.)
10
Executive Order 12642, “Designation of the Secretary of Defense as the Presidential
designee under title I of the Uniformed and Overseas Citizens Absentee Voting Act” (June 8,
16
Yet, presidents may not always fully, faithfully implement law. Figure 2.2 indicates that,
although these orders relate to the execution of law, the executive branch may seek to interpret laws
differently than Congress. We classify these as “adaptive” orders which implement the general
intent of legislation but the president may not implicitly follow the direction of the law or he may
use different statutes or even his constitutional authority to adapt policy to meet his objectives. For
instance, when President Reagan signed the Ethics in Government Act Amendments of 1985, the
law expressed the sense of the Congress that the government should “more aggressively use
suspension or debarment of contractors convicted of crimes.”11 Rather than acting pursuant to the
statute, President Reagan citing his Constitutional authority and the laws of the United States
generally, issued an executive order creating a government-wide system for debarment and
suspension from programs and establishing an Interagency Committee on Debarment and
Suspension to oversee the implementation of his order. 12 President Reagan followed the sense of
Congress but the order does not fall within the category of implementing law – instead, we consider
this as “adapting” law to meet the president’s policy objectives. Specifically, Congress did not
delegate authority to the president nor specify actions the government should take and unilateral
action, allowing him flexibility in developing his own policy.
Data, Methods and Expectations
While scholars generally classify all unilateral orders as “direct action” with the president
acting alone without Congress, in this chapter we segment and analyze our data set of unilateral
orders to identify a “direct action” order is issued at a specific moment in the policy process. By
definition, as explained in Chapter 2, the topic of a “command” order is on the president’s agenda
1988; 53 FR 21975). HR 4393, P.L. 99-410 Uniformed and Overseas Citizens Absentee Voting Act
(42 U.S.C. § 1973ff.
11
H. J. Res. 465, Ethics in Government Act Amendments of 1985 (Public Law 99-190, 99 Stat.
1185)
12
Executive Order 12549, “Debarment and Suspension” (February 18, 1986; 51 FR 6370).
17
which means it was discussed by him in a message, statement, address, speech or public address of
some type. In addition, the order has to come before any concurrent legislative action by Congress
is apparent. Routine orders are also direct action but they are distinguished from commands
because they are used to execute the president’s administrative duties, not the president’s agenda.
The lynchpin that separates a command order from a routine order is whether or not the issue in
question was a presidential promise or priority as defined by the president talking about it in public
and before Congress has acted. We operationalize direct action for the purpose of analysis by
creating a binary categorical dependent variable coded “1” for commands and “0” for routine
orders.
We restrict our analysis of orders to those that are in the “Executing Law” phase as shown
in Figure 1.1. This includes unilateral orders that implement or adapt a recently enacted law. We
determined that the president’s unilateral order was used to “faithfully” implement a law if the specific
statute (law) was cited in the order and if the law was enacted within one of the two legislative
sessions that make up the four years of the president’s administration. An order met the criteria for
“adapting” law if it related to a law enacted within one of the two legislative sessions that make up
the four years of the president’s administration but the president did not cite it exclusively as a source
of authority in the order. In order to determine if an order related to a recently enacted law, we
again analyzed the content of each order and then searched several electronic databases (discussed in
Chapter 2) in order to identify whether a law relating to the order had been enacted during the
president’s administration.13 A binary dependent variable was created and coded “1” if the unilateral
order was used to adapt law and “0” if it was used to implement law. The total number of orders
13
A legislative history was created for each unilateral order as part of this research. This enables us
to classify all the orders that met the criteria for “Direct Action” and “Legislative Process,” and
those that fell within the category of “Executing Law.” The next step was to determine which orders
were used to implement law and which were used to adapt law.
18
that related to the execution of recently enacted law includes a total of 482, of which 385 were issued
to implement law and 97 to adapt law.
The president commands when he acts first by setting policy unilaterally before Congress
can act. The advantage to the president is he does not necessarily “share” power with Congress. By
this classification, routine orders are not the subject of prior Congressional action. Our expectation
is that, in the battle for agenda control, the president will act when the issue is on the prior
Congressional agenda or the subject of current Congressional hearings because the president wants
to set policy before Congress does. We determine the extent of Congressional involvement in the
topic of the president’s unilateral action to understand when presidents are proactively acting first on
issues before legislation can be introduced. The topic of a unilateral order may have been the
subject of the prior Congressional agenda or current Congressional hearings. We investigate whether
the topic of each unilateral order was all or part of the subject of a Congressional committee hearing
before legislation has been introduced into the current legislative session (Congressional Hearing) or the
subject of a bill on Congress’s agenda in the prior session (Prior Congressional Agenda).
We also include several presidency-centered variables since the president has some degree
of control over the timing of his order with respect to political or temporal conditions. First, “Second
Half” is coded as the second half of a president’s term (either first or second term). Second, “Second
Term” is coded as the second term for a president (our data include Presidents Reagan, Clinton and
Bush as second term presidents). Related to these two variables, we interacted them to form a
variable (Second Half, Second Term). Third, “Election Year” is coded as any full year in which there is a
presidential election. These are consistent with how Mayer (1999; 2001) and Warber and Fine
(2012) construct these variables.
Following the logic from Chapter 4, we continue to explore the influence of the source of
authority presidents cite in their orders to justify their action. We expect that the president is more
19
likely to rely on statute when issuing routine orders and the Constitution for issuing a command
order. In the latter, he is executing a prearranged policy. In the former, he is acting more clearly
under his own prerogative. We include a binary categorical variable for orders where the president
issues an order citing a statute (Statutory Authority) and one for orders where the president cites his
Constitutional authority (Constitutional Authority). Unlike in Chapter 3 and 4, these sources of
authority include cases where statute was cited alone or in conjunction with other authority. We also
expand our measure of the presidential authority to include a variable measuring broad substantive
discretion (Substantive Discretion) with the expectation that greater discretion should yield a higher
probability of issuing a “command” order.
Several explanatory variables are also included to test our assertions pertaining to whether or
not presidents are working with or against Congress with respect to interbranch and intrabranch
gridlock. First, as a measure of intrabranch gridlock, or the measure of disagreement within a
chamber and or between the parties, we operationalized this in two ways. To create a measure of
internal distance between each of the two political parties in Congress (Party Distance) within each
chamber. This measure is a shorthand way of determining how likely it is that the parties in
Congress can get along enough to pass legislation; greater distances imply less cooperation while
closer distances imply more potential cooperation. We calculate the absolute value of the difference
of the CS-NOMINATE scores (which range from -1 for conservatives to +1 for liberals) for the
median member of each party (the average ideological Member) within the House and Senate (Poole
and Rosenthal 1991; 1997; McCarthy and Poole 1995).14 These measures represent an approximate
positioning of each institution on a liberal-conservative ideological scale. Next, to measure the
ideological distance within the parties in each chamber (Party Polarization) we subtract the absolute
14
Poole, Keith T., DW NOMINATE Data, website accessed, August 2011:
http://voteview.ucsd.edu/dwnl.htm .
20
value of the CS-NOMINATE scores for the extreme members within each party within the House
and Senate. In both cases, with Congress less able to act, as in instances where intrabranch conflict
is higher, presidents should issue more command orders since Congress will be less able to legislate
and since the president will want to act with dispatch.
Second, we measure interbranch gridlock between the president and Congress in three ways.
In the first measure we create a variable representing the ideological distance between the president
and the median member of each chamber (Distance between House / Senate Median and President). We
again operationalize this by using CS-NOMINATE scores for the president’s ideology and subtract
this from the chamber median of the majority party then take the absolute value of the difference
(see Deering and Maltzman 1999). In general, we should expect presidents to be more likely to issue
command orders when they are more likely to be at odds with Congress but more likely to issue a
routine order when they are more likely to be in agreement with Congress. We then operationalize
the size of the majority party by using a continuous variable of the total number of seats held by the
majority party in the House and Senate (Majority Party House/Senate). Presidents are likely to issue
more command orders when Congress is weaker and will issue fewer when Congress is stronger.
The third measure of interbranch gridlock measures the distance ideological between the president
and the filibuster pivot in the Senate (Filibuster Pivot). Specifically, we sort each Senator within each
session using the CS-NOMINATE scores and, taking the 60th member for each Congress, calculate
the absolute value of the distance between this member and the president.
Routine and Command Orders Over Time
To further explicate the use of routine and command orders in this phase of the policy
making process, Figure 1.4 charts the individual instances of the use of these orders by presidents.
Presidents appear to collectively use these orders in this policymaking phase in similar total amounts
over time. Two term presidents like Reagan, Clinton and George W. Bush all issued about 300
21
orders in this phase of the process, with Clinton and Bush using 307 and 309 respectively. One
term presidents like Ford, Carter and George H.W. Bush also used between 150 and 200 such
orders. The number of orders issued for Presidents Ford and Bush was only three orders different.
One implication to this commonality is that all presidents appear to have a consistent need to
manage government at this level using their unilateral power, consistent with the discussion above
about executive control and maintenance of government. Given that the orders in this category are
the most frequently used, this period is also the most important for study of the use of unilateral
executive power. Command orders tend to be more frequently used in eras where Congress is
institutionally stronger than the president. This is most acute in the Ford and Carter terms where
the number of command orders issued is lower than average for the presidents studied. President
George H.W. Bush, who served in a moment when Congress was resurgent following several
Reagan Administration scandals and a backlash against strong executive power from a Democraticcontrolled legislature. As presidential power swells during the Reagan, Clinton and W. Bush eras, so
too does the use of command orders. Perhaps not surprisingly, the use of command orders appears
to be highest when government is divided. This is especially true for the Reagan (149 command
orders) and Clinton (191) Administrations which made the most frequent use of these orders.
22
FIGURE 1.4
Use of Routine and Command Orders by President
250
200
150
100
50
0
Ford
Carter
Reagan
Routine
H.W. Bush
Command
Clinton
W. Bush
When Presidents Issue Routine and Command Orders
Because our dependent variable is a binary categorical one, we use a series of logistic
regression models, a generalized linear model for a binomial regression to test our hypotheses.15
This method allows us to predict a binary probabilistic response from a dichotomous variable based
15
These models allow us to (1) utilize a dichotomous variable arrangement (where yi = 1 if yi* > 0
and yi = 0 if yi* ≤ 0) and (2) determine the probabilistic effect of the key independent variables on
the dependent variable (Agresti 2002). That is, because the models require a dependent variable to
be binary, these models can test the above propositions regarding under what conditions a
preemptive order was issued. The model is briefly summarized as follows: Pr(y= 1 | x) = exp (α +
β x1 +…+ β xn ) / 1 + exp (α + β x1 +…+ β xn ), where xn is the total number of independent
variables described above. Since the estimated parameters from the equation does not provide
directly useful information, predicted probabilities (the predicted positive outcome for each
observation, given the values on the independent variables for that observation) and/or the marginal
change are used to estimate the relative importance of each coefficient (Long and Freese 2001).
Alternative specifications (such as an ordered logit and count models) revealed substantially similar
findings.
23
on the predictor (independent) variables. A positive coefficient indicates a president was more
likely to issue a “command” order while a negative coefficient indicates a “routine” order.16 The
results in Table 5.1 show the inclusion of several different models to estimate the institutional and
temporal effects on the use of unilateral orders as command or routine action. Presidents are more
likely to issue a command order when it relates to the subject of a Congressional hearing during the
current session which suggests that they act first to adopt policy in which Congress has an interest.
This finding demonstrates that the battle between the branches for the nation’s policy agenda is
partially adjudicated by the president’s use of use of information from Congressional hearings to
take unilateral action.
Presidents are also more likely to issue a command order when the issue was on Congress’s
prior legislative agenda. In the last model in Table 1.1, presidents are more than 32% likely to issue
a command order when the issue was on Congress’ agenda in the prior session. If the president
thinks Congress will consider the issue again, as evidenced by the fact that they considered it
previously, he will act first. For example, after being elected to a second term in office, President
Reagan issued Executive Order 12498, which signified his continued commitment to proactive
regulatory action directing agencies to prepare a yearly agenda containing all contemplated regulatory
actions for the coming year. (Mayer 2001, 127-128).17 This order was also effective in anticipating
regulatory reform legislation from the prior session.18 Congress may or may not approach the issue
16
Given the nature of these orders, it might be expected that those orders that dealt with defense,
emergencies, foreign aid or foreign trade would be more likely to engender a command order from
the president. This proved to be correct ―in models which truncated the sample to those issues,
the results were substantially similar. Because the results were fundamentally the same, we have
omitted these results for purposes of presentation.
17
EO 12498, Regulatory planning process (50 FR 1036; January 8, 1985).
18
Regulatory Reform Act: Hearings on H.R. 2327 Before the Subcommittee on Administrative Law
and Governmental Relations of the House Committee on the Judiciary, 98th Cong. 1st Sess. 312
(1983). S 1080 Regulatory Reform Act 97th Cong., 1st Sess., 127 CONG. REC. 7938 (1981); The
Senate passed S 1080 March 24, 1982 by a 94–0 vote. The Regulatory Reform Act Hearing on S.
24
in line with the president’s preferences, so, uncertain about the potential outcome, the president acts
first.
Other temporal variables in Table 1.1 are noteworthy. Presidents are more likely to issue a
command order in the second half of their terms (either first or second terms), suggesting presidents
act with some urgency in these moments. However, the joint effect of a president’s second half and
second term is not statistically significant, nor is the effect of the president’s second term.
Presidents are less likely to issue a command order in election years, although the effect is not
statistically significant. Presidents want to act presidential in election years but clearly find other
ways to demonstrate this than issuing unilateral orders. Each of these findings is robust across
different specifications (Tables 1.1 and 1.2). These findings are as curious for what they show as
much as for what they do not show. The only statistically significant relationship is that of the
president’s second half of his term ― none of the other time periods are statistically significant.
Although presidents are expected to act “presidential” or with greater dispatch in times when they
are in greater need for expedient action, these findings suggest that it is not always the case, even
considering that these orders are often used as a means to exercise “independent” power.
1080 Before the Subcommittee on Administrative Practice and Procedure of the Senate Committee
on the Judiciary 98th Cong., 1st Sess. 548 (1983).
25
TABLE 1.1
Interbranch Measures
Statue
Authority
Constitution
Authority
Substantive
Discretion
Prior
Congressional
Agenda
Second Term
.011
(.233)
-.277
(.234)
.015
(.236)
.087
(.238)
Second Half
.837 ***
(.185)
[2.7 %]
.665 ***
(.188)
[15.7 %]
.798 ***
(.182)
[21.3 %]
.891 ***
(.190)
[21.3 %]
Election Year
-.880 ***
(.198)
[-20.0 %]
-.818 ***
(.202)
[-18.1 %]
-.904 ***
(.184)
[-20.4 %]
-.781 ***
(.204)
[-17.9 %]
.106
(.314)
.192
(.316)
.068
(.321)
.015
(.323)
Congressional
Hearings
3.05 ***
(.155)
[64.1 %]
2.94 ***
(.158)
[62.6 %]
3.09 ***
(.158)
[64.5 %]
1.32 ***
(.170)
[62.2 %]
Statutory Authority
-.137 **
(.068)
[-13.1 %]
__
__
__
Constitutional
Authority
__
1.86 ***
(.268)
[34.7 %]
__
__
Substantive
Discretion
__
__
.920 ***
(.184)
[22.2 %]
__
Prior Congressional
Agenda
__
__
__
1.32 ***
(.170)
[31.8 %]
1,382
.298
-659.6
560.0 ***
53.7 %
1,382
.329
-630.0
619.3 ***
54.5 %
1,382
.310
-648.2
582.8 ***
53.7 %
1,382
.329
-630.5
618.2 ***
56.3 %
Second Half, Second
Term
N
Pseudo R2
Log-Likelihood
Likelihood Ratio χ 2
Reduction in Error
NOTE: Dependent variable: presidents issues a “command” order (dichotemous). *** indicates
statistical significance at p<.01 ** p<.05. * p<.10. Standard errors in parentheses. Percentage
expected change in brackets. Tests are two-tailed tests.
26
Models 2 and 3 examine related issues: the source of authority the president cites to justify
the order and the amount of discretion Congress has given them on that issue. In Model 2, the
source of authority as registered by statute is negative, suggesting that presidents are less likely (13%)
to cite a specific statutory basis of authority when they issue command orders. Put another way,
presidents are more likely to issue routine orders when they cite statute. This conforms to
expectations in that presidents, when issuing command orders, are generally fulfilling their role as
autonomous executive agents and are less concerned with accommodating Congress, at least with
respect to justifying their authority on an order. That is, when presidents go alone and issue
command orders as part of their political agenda, they are less likely to involve Congress, at least in
the text of the order. This corresponds to the findings from Chapter 4 and with other work that
suggests that presidents are less likely to cite Congressional authority when they have a political
incentive to do so (Bailey and Rottinghaus 2012). In Model 3, however, when the source of
authority cited is Constitutional, presidents are more likely to issue (40%) a command order in these
instances. Given the nature of the command orders, this is logical. Since these orders are
commonly used to subsidize executive power, presidents rely on an internal source of authority to
justify their actions.
The second to last column identifies models with delegated discretion. In general, greater
substantive discretion leads to a greater probability that presidents will issue a command rather than
a routine order (about 20% more likely). With broader discretion, the president has the added
authority to act in ways that may deviate from legislative intent. For instance, President George H.
W. Bush exercised his broad presidential discretion when ordering the selected reserve of the Armed
Forces to active duty in Iraq following the invasion of Kuwait.19 The conflict was the topic of
19
Executive Order 12727, 8/22/1990, Laws of the United States of America including Sections 121
and 673(b) of Title 10 of the U.S.C. 10 USC § 12304 - Selected Reserve and certain Individual Ready
27
committee hearings in both chambers of Congress regarding the emergency and necessary action,
yet the President acted first under broad “standing” discretion delegated under prior law.20
Presidents use their institutional unilateral authority, especially that authority that allows
greater use of discretion, to protect their “turf,” the executive branch. Presidents are often saddled
with a bureaucracy that they must attempt to control (Moe 1985; Moe 1998). As presidential
responsibility grew over time, presidents attempt to control the executive branch more furtively
(Dickinson 1997). As a result, presidents turn to formal and informal powers to navigate the
“administrative presidency” (Nathan 1983). Modern presidents use their unilateral toolkit to design
agencies which are responsive to their needs (Howell and Lewis 2002). As a means to encourage (or
demand) unity with executive goals, presidents may attempt to overcome vertical coordination
problems by using unilateral directives to induce compliance to presidential goals. Specific rules will
create coherent directives which allow for less interpretation and fewer vague calls for action (see
Aberbach and Rockman 2000). The clarity of the coordination and communication from the
president to the executive branch is important in facilitating program success (Krause and Dupay
2009), making presidential internal changes to the executive branch an important element of how
presidents manage their programs and attain policy success.
Table 1.2 also models interbranch variables pursuant to the expectations. In alternatively
specified count models (rather than the logit model presented here), the substantive effects are
Reserve members; order to active duty other than during war or national emergency House
Committee on Foreign Affairs Hearing on Sanctions Against Iraq, 8/1/1990, 8/2/1990, CIS NO
90-H381-105.
20
Senate Committee on Foreign Relations Hearing on U.S. Policy Toward Iraq: Human Rights,
Weapons Proliferation, and International Law, 6/15/1990, CIS No 91-S381-12; House Committee
on Foreign Affairs Developments in the Middle East, 6/20/1990, CIS No 91-H381-12, 7/31/1990
CIS No 91-H381-33; Joint Committee on Economic Hearing on Economic Outlook at Midyear
7/12, 8/2/1990 CIS No 91-J841-13; House Committee on Foreign Affairs Hearing on Sanctions
Against Iraq 8/1/1990, 8/2/1990 CIS No 90-H381-105, house Committee on Foreign Affairs
Hearing on Sanctions Against Iraq 8/1/1990, 8/2/1990, CIS NO 90-H381-105.
28
nearly identical (although the statistical significance is not always the same). The models here show
interesting divergent findings. In Models 1 and 2, when the ideological distance between the
president and either the House or Senate is greater, there is no statistically significant effect on
presidential action.21 However, in Model 3, there is a negative and statistically significant effect for
the size of the majority in Congress where presidents are more likely to issue routine orders when
the majority size is greater.22 This is contrary to expectations. Even so, the effects here are very
modest with a less than 1% change in probability of issuing a command order and of low statistical
significance.
Although the distance between party medians does not have a significant effect on increasing
the number of command orders, the greater distance from the filibuster pivot in the Senate does
have a significant effect on the probability the president will issue a command order. The effect is
substantively large with the president 61% more likely to issue a command order when the
ideological distance between the Senate filibuster pivot and the president is greater. An ideologically
closer filibuster pivot may allow the president to legislate more effectively and persuade the 60th
member of the Senate (and thus almost a supermajority of the Senate) to join the president’s policy
initiative. Greater distance implies less of a chance to convince an ideologically disparate Senator to
join the president’s coalition and thus cement some form of legislation acceptable to both. In this
case, presidents are more likely to issue a command order when the prospect of legislating seems
less likely.
21
In alternative models, we included both divided government and divided government interacted
with each of the interbranch conditions. Neither divided government nor the interacted terms were
statistically significant.
22
When run separately, the size of the majority party in the Senate had no substantive or statistically
significant effect while the majority party in the House had a small negative and statistically
significant effect. The pooled measure is therefore largely driven by the effect of the size of the
majority in the House.
29
TABLE 1,2
Interbranch Measures
Distance Between
President and Senate
Distance Between
President and
House
Majority Party Size
(Senate and House)
Senate Filibuster
Pivot
Second Term
-.070
(.252)
-.084
(.246)
-.070
(.252)
-.295
(.255)
Second Half
.852 ***
(.200)
[20.1 %]
.843 ***
(.198)
[19.8 %]
.852 ***
(.200)
[20.1 %]
.767 ***
(.199)
[18.1 %]
Election Year
-.700 ***
(.213)
[-15.7 %]
-.695 ***
(.213)
[-15.6 %]
-.700 ***
(.213)
[-15.7 %]
-.729 ***
(.214)
[-16.3 %]
-.001
(.334)
.041
(.342)
-.001
(.334)
.058
(.335)
Congressional Hearings
2.87 ***
(.162)
[61.5 %]
2.86 ***
(.166)
[61.4 %]
2.87 ***
(.162)
[61.5 %]
2.87 ***
(.167)
[61.5 %]
Constitutional Authority
1.91 ***
(.271)
[35.5 %]
1.91 ***
(.271)
[35.5 %]
1.91 ***
(.271)
[35.5 %]
1.83 ***
(.275)
[34.4 %]
Substantive Discretion
.974 ***
(.188)
[23.2 %]
.976 ***
(.188)
[23.2 %]
.974 ***
(.188)
[23.2 %]
1.01 ***
(.190)
[24.2 %]
Prior Congressional Agenda
1.28 ***
(.178)
[31.1 %]
1.29 ***
(.178)
[31.1 %]
1.28 ***
(.178)
[31.1 %]
1.31 ***
(.179)
[31.6 %]
Distance Between President and
Senate
-.238
(.446)
__
__
__
Distance Between President and
House
__
-.360
(.487)
__
__
Majority Party Size (Senate and
House)
__
__
-.021 *
(.014)
[-.5 % ]
__
Senate Filibuster Pivot
__
__
__
2.56 ***
(.782)
[61.0 % ]
N
Pseudo R2
Log-Likelihood
Likelihood Ratio χ 2
Reduction in Error
1,382
.376
-586.0
707.2 ***
57.1 %
1,382
.376
-585.9
707.4 ***
57.1 %
1,382
.376
-586.0
707.2 ***
57.1 %
1,382
.381
-580.8
717.7 ***
55.4 %
Second Half, Second Term
____________________________________________________________________________________________
NOTE: Dependent variable: presidents issues a “command” order (dichotemous). *** indicates
statistical significance at p<.01 ** p<.05. * p<.10. Standard errors in parentheses. Percentage
expected change in brackets. Tests are two-tailed tests.
30
Table 1.3 investigates the intrabranch conflict (where Congress is internally divided).
Interestingly, presidents are less likely to issue a command order when the ideological distance
within the parties (party extremes) in each chamber is greater, although the effect is only modestly
statistically significant at the p<.10 level and is only apparent in the Senate. This contradicts
expectations from Chapter 2 where we thought that presidents would take advantage of Congress’
presumed inability to act by issuing more command orders. Presidents do not want to step on the
toes of a dysfunctional party which could still be capable of making law or challenging the president.
Yet, examining a related but distinct environment, presidents are more likely to issue a command
order when the ideological distance between the median members between each party in each chamber
is greater. This finding is true for both the House (47% more likely) and the Senate (73% more
likely). Thus, the greater the inability of the parties within each branch to reach agreement (based
upon greater ideological disparities between them), and thus form coalitions, the more likely the
president is to issue a command order. Put together, presidents are less likely to issue a command
order when the parties are internally divided but more likely when the party medians in each chamber
are farther apart.
31
TABLE 1.3
Intrabranch Measures
Distance Between
Party Extremes
(House)
Distance Between
Party Extremes
(Senate)
Distance Between
Party Medians in
Senate
Distance Between
Party Medians in
House
Second Term
-.095
(.246)
-.152
(.248)
-.238
(.256)
-.231
(.261)
Second Half
.820 ***
(.205)
[19.3 %]
.847 ***
(.198)
[19.9 %]
.852 ***
(.198)
[20.1 %]
.811 ***
(.198)
[19.1 %]
Election Year
-.694 ***
(.213)
[-15.6 %]
-.711 ***
(.213)
[-16.0 %]
-.458 ***
(.211)
[-15.6 %]
-.698 ***
(.213)
[-15.7 %]
-.029
(.338)
-.139
(.340)
-.048
(.333)
-.005
(.332)
Congressional
Hearings
2.86 ***
(.167)
[61.4 %]
2.87 ***
(.167)
[61.5 %]
2.85 ***
(.166)
[61.2 %]
2.85 ***
(.166)
[61.3 %]
Constitutional
Authority
1.89 ***
(.276)
[35.5 %]
1.76 ***
(.285)
[33.5 %]
1.76 ***
(.282)
[33.5 %]
1.79 ***
(.282)
[34.0 %]
Substantive
Discretion
.978 ***
(.188)
[23.3 %]
1.00 ***
(.189)
[23.8 %]
.955 ***
(.178)
[22.7 %]
.962 ***
(.188)
[22.9 %]
Prior Congressional
Agenda
1.28 ***
(.178)
[30.8 %]
1.26 ***
(.178)
[30.6 %]
1.26 ***
(.178)
[30.6 %]
1.26 ***
(.178)
[30.5 %]
Distance Between
Parties in Chamber
-.166
(.836)
-.438 *
(.263)
[-10.4 %]
__
__
Distance Between
Chamber Median and
President
__
__
1.97 *
(1.04)
[47.1 % ]
3.07 *
(2.08)
[73.2 % ]
N
Pseudo R2
Log-Likelihood
Likelihood Ratio χ 2
Reduction in Error
1,382
.376
-586.2
706.9 ***
57.1 %
1,382
.377
-584.8
709.7 ***
56.1 %
1,382
.378
-584.4
710.4 ***
58.2 %
1,382
.377
-585.1
709.0 ***
58.0 %
Second Half, Second
Term
NOTE: Dependent variable: presidents issues a “command” order (dichotomous). *** indicates
statistical significance at p<.01 ** p<.05. * p<.10. Standard errors in parentheses. Percentage
expected change in brackets. Tests are two-tailed tests.
32
Faithful or Adaptive Implementation
In Figure 1.5 we see the dual roles of the “administrator” and “independent” executive at
work as subdivided by individual administrations. Several findings emerge from this graph. First,
although there is variation in the use of unilateral orders at this execution of law stage, each of the
presidents issued unilateral orders more frequently to implement rather than adapt law. The use of
unilateral orders to implement policy change is always greater than the number of instances
adaptation. As noted earlier, the presumption in this stage of the process is for implementation to
be faithful. Second, there is a steady rise in the number of instances of orders being used to
“adapt” laws beginning with the Ford Administration and rising to a peak during the Clinton
Administration. Divided government appears to be a culprit here where Presidents Reagan and
Clinton both issued more orders which adapted legislation to their own ends. The frequent uses of
administrative actions such as these to control the executive branch were especially prominent
during these two Administrations (Kagan 2001).
33
FIGURE 1.5
Presidents and the Execution of Law
120
100
80
60
40
20
0
Ford
Carter
Reagan
Implement
Bush
Clinton
W. Bush
Adapt
When Presidents Implement or Adapt Legislation
Table 1.4 displays the results for the logistic regression model intended to show the
difference in likelihood that a president will issue an order adapting law over one that implements
law. Positive coefficients signal coefficients that are more likely to adapt legislation while negative
coefficients signal coefficients more likely to implement legislation. Because of the high correlation
between ideological distance between the President and the Senate filibuster pivot and the
ideological distance between the President and the House median member, the two variables were
estimated in separate models. As expected, when presidents decide to “adapt” legislation using a
unilateral order, we see the picture of a more independent executive. First, he is more likely to cite
Constitutional authority (14% to 16% more likely) when issuing an “adaptive” order. Second,
presidents are more likely to adapt legislation when the issue is on his agenda. This happens when
Congress is pursuing one policy direction and the president another, as the example concerning
34
President Carter and the management and destruction of nuclear waste described above
demonstrates. Third, presidents are also more likely to issue a “faithful” order when they have
submitted a legislative initiative to Congress related to the law (8% to 9% more likely). It is more
likely that presidents have greater input into the content of the law through the legislative process,
therefore minimizing their need to “adapt” the legislation to their needs.
Fourth, when presidents have greater substantive delegated discretion in the law that they are
tasked with implementing, they are less likely to faithfully implement and more likely to adapt the
law. As Congress provides the executive with greater authority to act, presidents take advantage of
this authority to alter the way in which the law is executed. However, the effect is modestly
substantively (at about 5% more likely) and statistically significant and not robust across all models
in Table 1.4. The president may need to initiate policy action as well as effectuate change within the
executive branch. While Congress will describe the policy action, it generally does not always specify
how the president is to manage internal changes within his own branch of government, which leaves
the president more room for interpretation.
Yet, we also find, also contradicting expectations, presidents issue more “faithful”
implementation orders in the second terms or in the second halves of their term in office
(consistently between 10% and 18% more orders). The coefficient for these variables is negative in
each model, signaling greater probability of faithful implementation. Presidents appear to issue
more “faithful” implementation orders when legislative sessions are coming to a close, such as in the
second half of the president’s term and in the second term. This suggests that the president is not
the lame-duck he is often accused of being and that the two branches are effectively able to work
together even as their collective time in office dwindles. It may alternatively be that presidents are
less willing to run the risk of angering a Congress with whom they may have additional business with
in the subsequent sessions. In either case, this finding defies expectations from Chapter 2 which
35
indicate that these should be moments where the president acts more independently to get his way.
This finding suggests that these orders may be consistent with the president’s role as
“administrator.”
However, when jointly in the second half of the second term, the true definition of “last
minutes” action, the likelihood of issuing an adaptive order is much greater (22% to 27% more
likely). The implication is that when presidents are at the very end of their tenure in office they are
much more likely to use their orders to facilitate their own policy as a means to leave a legacy when
leaving office. Presidents work diligently to ensure that their policy accomplishments last. Late in
his second term in office, President Obama initiated discussion on rule imposing new fuel efficiency
and greenhouse gas standards for big rigs and semi-trucks. The new regulations, “to be drafted by
the Environmental Protection Agency and the Transportation Department by March 2015 and
completed a year later so they are in place before Mr. Obama leaves office, are the latest in a series
of actions intended to cut back on greenhouse gases without the sort of comprehensive legislation
the president failed to push through Congress in his first term” (Baker and Davenport 2014).
Indeed, the authors conclude, “the biggest obstacle facing Obama’s regulatory agenda isn’t the
Republican House, or the opposition of business groups, it’s the calendar” (Goad 2014).23
23
Accessed at: http://thehill.com/blogs/regwatch/pending-regs/198999-obamas-rush-toregulate#ixzz2uG56dN4i
36
TABLE 1.5
Logistic Regression Models for Adapting Law, Interbranch Measures
______________________________________________________________________________
Divided Government
Distance Between
President and Senate
Distance Between
President and House
Distance Between
President and Senate
Filibuster
Second Term
-1.83 ***
(.494)
[-17.5 %]
-1.91 ***
(.513)
[-18.3 %]
-1.66 ***
(.483)
[-16.5 %]
-1.61 ***
(.444)
[-15.9 %]
Second Half
-.987 ***
(.374)
[-12.3 %]
-1.02 ***
(.382)
[-12.9 %]
-.888 ***
(.369)
[-11.2 %]
-.864 ***
(.366)
[-10.9 %]
Election Year
.637 **
(.344)
[7.9 %]
.611 **
(.342)
[7.7 %]
.557 *
(.338)
[7.0 %]
.528 **
(.338)
[6.5 %]
Second Half, Second Term
1.54 ***
(.600)
[24.7 %]
1.65 ***
(.601)
[27.2 %]
1.47 ***
(.601)
[23.8 %]
1.46 ***
(.600)
[23.7 %]
Constitutional Source of
Authority
1.00 **
(.479)
[9.2 %]
.953 **
(.480)
[8.9 %]
1.03 **
(.478)
[9.6 %]
1.02 **
(.477)
[9.5 %]
Substantive Discretion
.503 *
(.298)
[5.8 %]
.471 *
(.295)
[5.4 %]
.435 *
(.292)
[5.1 %]
.448 *
(.290)
[5.2 %]
Presidential Agenda
2.21 ***
(.396)
[22.4 %]
2.19 ***
(.396)
[22.4 %]
2.09 ***
(.387)
[21.7 %]
2.03 **
(.383)
[21.0 %]
Presidential Legislative Initiative
-.679 **
(.307)
[-6.9 %]
-.626 **
(.306)
[-6.5 %]
-.645 **
(.304)
[-6.8 %]
-.725 **
(.306)
[-7.5 %]
Divided Government
.759 ***
(.308)
[8.3 %]
__
__
__
Distance Between President and
Chamber
__
1.58 **
(.792)
[18.4%]
1.12
(.849)
__
Distance Between President and
Senate Filibuster
__
__
__
2.60 **
(1.42)
[30.5 %]
475
.157
-194.2
72.7 ***
6.67 %
475
.152
-195.3
70.5 ***
1.11 %
475
.147
-196.4
68.2 ***
3.33 %
475
.151
-195.7
69.7 ***
5.56 %
N
Pseudo R2
Log-Likelihood
Likelihood Ratio χ 2
Reduction in Error
_____________________________________________________________________________________
NOTE: Dependent variable: coded 1 for orders that adapt law and 0 for orders that implement law. ***
indicates statistical significance at p<.01 ** indicates statistical significance at p<.05. * indicates statistical
significance at p<.10.
37
Table 1.5 and 1.6 also specifically consider various specifications of interbranch and
interbranch relations. First, in most all of the specifications measuring probable interbranch
differences, presidents are more likely to issue orders that adapt legislation. For instance, Presidents
are also more likely to issue an adaptive order when government is divided (8% more likely). The
effect is greater when specifically considering the Senate. Presidents are more likely to challenge
Congress by issuing orders that adapt law as the ideological distance between the president and the
median member of the Senate increases (18% more likely) and the Senate filibuster pivot increases
(30% more likely). Second, for interbranch relations (Table 1.5) we examine the ideological distance
between the two parties within each chamber. With an increase in polarization in the House and the
Senate (ideological distance between the median members of each party), the president is also
significantly more likely to issue an adaptive order. The effect is almost twice as large in the House
(80% more likely) as in the Senate (45% more likely).
38
TABLE 1.6
Logistic Regression Models for Adapting Law, Intrabranch Measures
______________________________________________________________________________
Distance Senate
Party Medians
Distance House
Party Medians
Second Term
-1.77 ***
(.496)
[-17.3 %]
-1.82 ***
(.509)
[-17.7 %]
Second Half
-.778 **
(.361)
[-9.7 %]
-.891 ***
(.367)
[-11.2 %]
Election Year
.530
(.336)
.520
(.336)
Second Half, Second Term
1.53 ***
(.600)
[24.7 %]
1.65 ***
(.603)
[27.4 %]
Constitutional Source of
Authority
.894 **
(.488)
[8.5 %]
.939 **
(.486)
[8.9 %]
.449
(.291)
.446
(.290)
Presidential Agenda
1.98 ***
(.383)
[20.5 %]
1.98 ***
(.383)
[20.6 %]
Presidential Legislative
Initiative
-.679 **
(.303)
[-7.0 %]
-.671 **
(.302)
[-7.0 %]
Distance Between Chamber
Medians
3.82 **
(1.92)
[44.8 %]
6.80 **
(3.83)
[79.9 %]
N
Pseudo R2
Log-Likelihood
Likelihood Ratio χ 2
Reduction in Error
475
.152
-195.3
70.4 ***
4.44 %
475
.151
-195.7
69.6 ***
2.22 %
Substantive Discretion
_____________________________________________________________________________________
NOTE: Dependent variable: coded 1 for orders that adapt law and 0 for orders that implement law. ***
indicates statistical significance at p<.01 ** indicates statistical significance at p<.05. * indicates statistical
significance at p<.10.
39
These findings typify the argument in this paper. Presidents use the leverage they have to
pursue their agenda using their unilateral orders. In the present case, when presidents implement
legislation into law, this is the last chance they have in the process to bend the outcome to their
preferred policy agenda. Moments when presidents find a political antecedent in Congress (in
various forms), presidents are more likely to adapt legislation to their preferences. This adaptation
does not necessarily mean a rewriting of the legislation or that presidents can alter the full impact of
the legislation. Even if this adaptation process allows presidents to minimally alter the fate or
function of a policy, presidents retain a significant advantage over Congress. Indeed, more than any
of the other types of aggressive, “independent” presidential unilateral action (command orders from
above), the president’s adapt orders described in this chapter are most likely to be used during
moments of significant interbranch friction, especially in the Senate.
Conclusion
The findings with respect to unilateral orders show that political and institutional factors
influence the president’s decision to issue an order to implement law. First, presidents act as
“administrators” in several scenarios. Presidents are more likely to implement faithfully when the
legislation was of their initiative. This is not surprising because presidents are able to get more of
what they want when they initiate the process. Second, as scholars have demonstrated, there is an
institutional restriction at play where presidents are more likely to issue orders that implement law
when their discretion is narrowed by Congress. The president may decide against pushing his
administrative agenda instead, he continues to keep the wheels of government turning. Presidents
“faithfully” implement laws more consistently in their second half of terms and second terms (the
end of legislative sessions). Presidents again appear to respect the balance of power and the need to
carry through legislative action more is greater than their desire to push their own agenda. The
40
Constitution imbued each lawmaking institution with a different set of responsibilities, together
possessing overlapping and interconnected authority.
Meanwhile, presidents also act as “independent” executives in moments when their desire to
push their agenda is greater when ideological differences suggest Congress may not be willing to
support them. We find that the president is more likely to issue orders that adapt law relying on
broad substantive discretion and constitutional authority. Presidents also adapt legislation when
party polarization is higher or when the distance between the presidents and the Senate filibuster
pivot is larger. Presidents also ensure their agenda is more likely to be met in the implementation of
legislation by being more likely to adapt laws on their agenda and more likely to faithfully implement
laws when they initiated the legislation. These actions taken together demonstrate panoply of
interests when presidents take advantage of their first and last mover advantage. Congress is
demonstrably unable to stop the president at this stage giving the president a greater hand in
executing his preferences.
41