Exercising Unilateral Discretion: Presidential Justifications of Unilateral Powers in a Shared Powers System

Delegation and Interbranch Relations

The president’s unilateral powers are linked to the type of justification they cite in each order as the basis for their authority to act. The legal justification employed by a president is often delegated by Congress. Delegation is nested within the principal–agent relationship where principals (Congress) delegate authority to agents (the president or the executive branch) to develop and implement policy (Bendor & Meirowitz, 2004; Mitnick, 1973, 1975; Waterman & Meier, 1998). Under optimal conditions, by delegating its authority to those with knowledge and expertise, Congress reduces high levels of technical uncertainty and information asymmetry and, in exchange, it receives enhanced information concerning policy alternatives and outcomes (Bawn, 1995; Epstein & O’Halloran, 1994; Gilligan & Krehbiel, 1987; Kiewiet & McCubbins, 1991; McCubbins, 1985). Delegation also enables Congress to absolve itself of the costs of policymaking, which are borne instead by the president and the executive branch, satisfying both the president and Congress. Therefore, Congress delegates broad or narrow authority, depending on their interests, and its will is ultimately is reflected in the nature of the design of the legislative controls (Moe, 1987).

By delegating broadly, members of Congress can claim credit for the benefits of policy outcomes without having to bear the costs as decision making is borne by the president who has the freedom to act first and decisively (Epstein & O’Halloran, 1999; Fiorina, 1982). Fiorina (1986) refers to this as “shifting responsibility,” where Congress delegates policy implementation specifically to the executive to either avoid blame or because of the policy expertise of the executive. It can also be a strategic move on the part of Congress to gain an advantage over the executive by giving them an opportunity to act as the first mover (Bendor, Glazer, & Hammond, 2001). However, narrowed discretion may confer detailed instructions that constrain the executive branch to produce policy outcomes consistent with Congressional objectives, leaving the executive branch with little room for independent action (Bawn, 1997; Fleishman & Aufses, 1976; McCubbins, Noll, & Weingast, 1987, 1989; Moe, 1990). Presidents employ discretion in a unilateral order as a means to maximize their policy influence—This occurs within the boundaries of what Congress establishes by statute. As Shugart and Carey (1992) put it, an executive cannot exceed the authority given to it by the legislature. Ultimately, the courts arbitrate the credible degree of delegation allowable (as directed by Congress) and the proper use of it (as executed by the president).

Delegated Power, Legal Constraints, and Independent Action

The degree of independence of executive action is important in the study of unilateral power because it allows for assessment of the possible encroachment and limitations of shared powers between the branches through delegation. To fully develop our understanding of how presidents act, we need to understand the influence of discretion on the president’s use of authority (Fenno, 1958). As Mayer (2009) argues, “we might make more progress by looking at what happens before a president takes unilateral action, rather than after” (p. 448). Indeed, presidents are required to justify their actions in some direct manner, and the White House of Office of Legal Counsel and the Attorney General have the formal responsibility to ensure a legal pathway for justification (Gibson, 2008). ² Presidents most frequently justify unilateral action based on Congressional authority (Rudalevige, 2005; J. D. Bailey & Rottinghaus, 2014) but also rely on other sources. Congress may appear passive and the president strong but this view minimizes the role of Congressional delegation of power, including the ability to retract that delegation if necessary (Moe, 1998; Shugart & Carey, 1992). Yet, the literature has neither fully unpacked the influence of delegated discretion on the president’s exercise of unilateral powers nor the putative political restraint of the Court.

Evidence from internal White House suggests that presidents are selective about what statutes with which they justify their authority in a unilateral order. For instance, in an internal White House memoranda in the Carter Administration objecting to draft language of an executive order that would have required agencies to consider the environmental effects of federal actions in other countries, the Department of Energy (DOE) objected because the executive order could be read as tethered to Congressional authority rather than the president’s power. Specifically, the DOE review counsel recommended that the order be structured to avoid one source of authority in favor of another so the decisions rest “expressly on the inherent authority of the President and not tied to” the National Environmental Policy Act. ³ Presidents and their advisors work to operate within the boundaries of the law, a dilemma President Obama faced when 2014 advocates of immigration reform called on him to make changes in deportation policy, referenced at the beginning of this article. President Obama hesitated but subsequently decided to take executive action by locating a series of authority sources that allowed the White House to justify their actions—The Department of Justice issued a 33-page legal memorandum on the president’s authority to do so that outlined the options they believed they had to choose among certain statutes (Shear, 2014).

The influence of discretion on the use of unilateral power needs additional scrutiny as it informs us about the boundary of possible unilateral presidential action. Indeed, Howell (2003) argues that the president’s exercise of “powers delegated to them by different congresses, past and present, and other times by reading new executive authority into the Constitution itself” (p. 13). ⁴ The overarching assumption in theories of delegation, and those centered on unilateral action, is that when discretion is delegated broadly, the executive branch is able to act with greater independence from Congress in adopting final policy outcomes. However, discretion is not always delegated broadly. Instead, there is variation in the extent of authority and the type of constraints Congress delegates to the president and this may have a discreet influence on the president’s exercise of unilateral action. Similarly, the Court may reject overly broad claims of authority (Liptak, 2014). Presidents balance the scope of their orders, consider the issues in question, and legally justify their actions with the hope of executing the order and maintaining a strategic institutional balance (Belco & Rottinghaus, 2017). Therefore, the goal in this article is to expand into an area the literature leaves unanswered, which is the question of whether discretion delegated (broad or narrow) by Congress and a possible reaction from the Court influence the president’s propensity to act with broad unilateral action.

Joint Institutional Friction and Pushback on Unilateral Power

For unilateral powers, the three branches engage in a fluid, complex and ongoing dialogue about the proper scope of executive authority (Cooper, 2002). When the courts rule against the president, they most often do so on Congress’ behalf, claiming that Congress had a specific intent in mind (Melnick, 2005). The Court is the only arbiter of the constitutionality of the practice of delegation by Congress (Schoenbrod, 1993). Yet, even when a president has been granted greater discretion, presidential actions are still checked by the Courts who may temper the use of executive authority (O’Donnell, 1994). This applies even to cases where presidents justify their actions based upon specific statutes where the Court may reason that the law has been misapplied (Whittington, 2009). In order for a unilateral action to be considered valid, the Court generally requires that (a) the statue not be in violation of the Constitution, (b) the statute be within the proper Constitutional authority of Congress, and (c) the statue does not violate the “non-delegation” doctrine (the delegation from Congress must be reasonable and/or provide the ability for unfettered action to the president; Cooper, 2002, p. 22; Epstein & O’Halloran, 1999, p. 20).

Presidents gain legal authority to act from both the Constitution and from statutes, making the Court a key element in the explanation of the use of presidential prerogative (Whittington, 2009). This arrangement is supported by Justice Robert Jackson’s famous “practical grouping” of presidential power in Youngstown Sheet and Tube Co. et al. v. Sawyer, 343 U.S. 579 (1952). When the president acts under authority delegated by Congress, in addition to his constitutional authority, presidential power is at its maximum and most safe from judicial oversight. This is confirmed by recent scholarship, which has shown that even during wartime the Supreme Court usually casts itself as an auditor of process, checking to see if Congress has delegated authority to the president, rather than as an umpire of zero sum contests between executive power and individual rights (Pildes & Issacharoff, 2004). Because, as Justice Jackson wrote in Youngstown, “presidential powers are not fixed but fluctuate, depending on their disjunction or conjunction with those of Congress,” the Court may intervene to limit the exercise of unilateral power on Congress’ behalf. This intervention is rare in practice but this arrangement does provide for the possibility of judicial oversight and this may alter the amount of discretion employed by the president in a unilateral order. Joint, Congress–Court rejection of a president’s orders would limit executive reach so presidents are strategic in considering institutional disagreement when they face a united front in the other two branches of government.

Expectations of Unilateral Orders Under Delegated Discretion

We presume the circumstances under which the president will issue a unilateral order with broad or narrow discretion are influenced by the political constraints of Congress and the Courts. Mayer (2009) argues that “broad grants [of authority], ambiguity and the potential for swift action are a compelling combination” (p. 432). When presidents have broad discretion, it gives the president the authority to act when they determine such action to be expedient. Presidents in general prefer to justify their orders with broad discretion, providing them maximum flexibility and the greatest latitude to secure their policy aims with an order. Yet, there is a potential cost to issuing an order with broad authority, even if properly used, in that Congress may object to an expansive use of authority or the courts may strike the order down on the grounds it exceeds the executive’s authority. A Congress and Supreme Court which is at ideological loggerheads with a president may object to the president’s use of an order with broad discretion because of the policy latitude it affords a president. A president must weigh the possible backlash in their justification of the authority relied upon in an order. Although unilateral orders are rarely overturned, the prospect of a rebuke from either Congress or the Courts (and risking permanent inability to act in the future) may alter a president’s behavior.

Strategic presidents will issue an order with broad justification in political circumstances when they think they can get away with it. What political circumstances motivate presidents to cite broad authority? We argue the greater the distance in ideology between the president and the median member of Congress, the more likely the president should be to issue more orders under broad discretion. When the president and Congress are similar in ideological orientation, presidents are more comfortable broadening their legislative reach through statute with a unilateral order because they believe Congress will back them up if there are any interpretive differences. Congress is comfortable allowing (not challenging) the president who issues an order with broad discretion because they generally agree ideologically and they trust the president to pursue an agreeable policy outcome. From these findings and extending this logic, we estimate the distance between the president and the median member of Congress and expect that presidents are more likely to issue an order with broad discretion when there is greater distance in ideology between the president and the median members of each chamber (Hypothesis 1).

Using similar logic, the greater the distance in ideology between the president and the median member of the Supreme Court, the more likely the president is to issue more orders with broad discretion (independent of Congress; Hypothesis 2). With broad discretion, the president is comfortable exercising unilateral power when only one branch may push back. The Court has also usually avoided solo, head-on confrontations with the chief executive (Cooper, 2002). The Courts have generally been deferential to the executive, especially on matters involving executive prerogative, in times of crisis and when the president has taken decisive action (Melnick, 2005). The president has flexibility to execute the laws but this ends if he seeks to alter the “plain letter of the law” (Rudalevige, 2014). The Court’s doctrine of decision making on cases involving executive power favors the president and, even when challenged, presidents are more likely to reap outcomes favorable to their use of executive power (specifically challenges to the use of executive orders) when the majority of the justices deciding the case are from the same political party as the president (Howell, 2003). Ideological goals within the Court that are more proximate from the White House’s ideology (often on issues of executive power) yield a president who should be more willing to exercise the full weight of his discretion in unilateral orders and a court that embraces a doctrine to allow more executive institutional control.

The additive frictional effect, however, should have the opposite outcome. Presidents want to avoid joint confrontation with both Congress and the Courts given that this is a powerful recipe for halting executive use of power (either temporarily or permanently) as discussed in the previous section. For instance, when the executive is delegated broad discretion, it positions him to implement policy that may deviate from legislative intent (Mayer, 2009) prompting a possible judicial challenge (Thrower, 2016). Because the president can justify his order using a number of statutory and nonstatutory options, presidents tend to choose an option that is least odious to possible judicial oversight (J. D. Bailey & Rottinghaus, 2014; Quint, 1989). Wishing to avoid an interbranch entanglement, presidents may “seek favorable legislative terms that can insulate their decisions from judicial review” (Whittington, 2009, p. 655). Being generally risk averse, in practice this implies that presidents mitigate larger claims of discretionary action when acting unilaterally. As Justice Jackson’s holding implies, presidents who run afoul of Congress (here operationalized as greater ideological distance) and who face a court predisposed to disagree with a president (here also operationalized as greater ideological distance), a strategic president should adjust their citation pattern and narrow their authority as cited in the text of the order.

Therefore, when the president’s ideology is farther from both the median justice’s ideology and the median member of a Congressional chamber’s ideology, he should (a) continue to use their discretionary unilateral power but (b) mute that power so as not to aggravate two branches of government (Hypothesis 3). ⁵ That is, we should continue to see presidents utilizing executive power based on statutory authority, but that power will be justified using more narrow authority instead of broad authority. When the president and Congress are farther apart ideologically, but not the Court, presidents face the need for greater interpretative room to establish an executive order and will maximize their ability to control the outcome of that policymaking process. The Court may decline to intervene and/or Congress may ignore the issue, but the potential of disagreement or stoppage of a policy should be enough incentive for presidents to self moderate their use of discretion. The potential of having two branches of government jointly minimize a policy and set a precedent that may be difficult to work around will incentivize a strategic president to narrow the scope of their authority under the cited justifications in the text of the order. This allows the White House to insulate their unilateral decision-making process for possible review and rejection.

Empirical Strategy

We have coded all executive orders issued by presidents from 1951 to 2009. ⁶ We selected executive orders because they are the most frequently used and significant unilateral orders found over time (Cooper, 2002; Relyea, 2007). Executive orders, as Mayer (2001) writes, are “presidential directives that require or authorize some action within the executive branch (though they often extend far beyond the government)” (p. 4; see also Warber, 2006). To collect these executive orders, several sources had to be searched to generate this complete record including, the Codification of Presidential Proclamations Disposition Tables, the United States Statues at Large, the United States Federal Code, the Congressional Information Service (CIS) Index to Presidential Executive Orders and Proclamations, the United States Federal Register, and the Public Papers of the Presidents. The combined search yielded a comprehensive list of 2,510 executive orders during the 1951 to 2009 period. Because we are interested in the likelihood of issuing an order using certain types of discretion, we use each order as the unit of analysis rather than a count model of the number of orders under each type of discretion. In addition, to extend our analysis to examine whether significant statute and nonstatute orders are statistically unique, we undertook an analysis to test this question. A search was made to match the data to Mayer’s (1999) and Howell’s (2003) span of executive order data and to allow for the 15-year “significant” cutoff as identified by Howell (2003). There were a total of 224 significant orders by this measure.

We created two dependent variables to capture the measure of discretion delegated by Congress. Although scholars have developed different methods for measuring discretion, as noted above, we measure discretion based on the key statute cited in the executive order. The relevant portions of the statues and changes were then compiled from the United States Code and the U.S. Statutes at Large, to ensure we captured the law as it was for the correct time period. If multiple statutory citations were included in the order, we coded the controlling legal citation that was most central to the function of the order. We then coded each relevant statutory provision using two types of discretion: (a) substantive and (b) ex-ante procedural control. The intention was to analyze more fully the type and extent of statutory constraints placed on a president’s ability to exercise discretion. ⁷ The online appendix uploaded on the American Politics Research (APR) website has a representative list of statutes and associated levels of discretion that were coded from each order. Intercoder reliability checks demonstrate significant intercoder agreement. ⁸ The online appendix includes the material for the reproduction of the discretion variables. The number of orders that cited statutory authority (and that therefore could be matched to a degree of discretion) totaled 2,258 during the 1951 to 2009 period. ⁹ For significant orders, 210 matched orders were located.

The primary assertion for the dependent variable is that presidents may choose among several alternatives to justify their unilateral use of power—Some statutes offer greater flexibility or authority, whereas others offer less. As presidents shop for the appropriate legal justification, they may choose an option that either maximizes or minimizes their authority, depending on the prevailing political conditions. Issuing an order relying on broad discretion is appealing to presidents because they are able to accomplish more, retain specific authority to interpret the scope of an order, and may have more significant legal grounds on which to appeal. But, by issuing broad order, presidents may risk backlash from Congress or the Courts, perhaps by legally halting the effect of the order or legislatively narrowing the president’s ability to act in the future. Issuing an order with narrow discretion is less costly for the president in terms of interbranch disagreement but the effect of such orders may be modest in scope.

For the first type of discretion, we proceeded by analyzing the general substantive limits placed on the president based on three categories: narrow, limited, and broad discretion. The president had narrow discretion when simply fulfilling orders from Congress with no flexibility to modify or interpret the law. An example is Executive Order 11738 where Congress required the executive to establish procedures and guidelines to implement statutes from specific legislation. The Order notes that the action “will result in effective enforcement of the Clean Air Act (hereinafter referred to as ‘the Air Act’) and the Federal Water Pollution Control Act (hereinafter referred to as ‘the Water Act’).” ¹⁰ Next, the president had limited discretion if the language of the statute required him to comply with conditions or to follow directions but the president had some discretion over whether or not to declare a specific policy or its content. An example of this would be Executive Order 12483, where, under Section 5 of the Trade Act of 1973, the President may include or exclude specific countries, goods, or tariffs but only if certain conditions apply. ¹¹ Finally, the president was considered to have broad discretion if there were no express limits on the delegating authority. An example of this would be Executive Order 12468 where, under the Federal Advisory Committee Act, the president has complete discretion to establish an advisory committee with no significant restrictions. ¹²

The second type of discretion we analyze captures ex-ante procedural controls placed on the president (Spence, 1999). These are generally administrative procedures that constrain the president’s exercise of discretion as part of the decision-making process. The primary means used require or recommend consultation with an advisory commission, secretary, or agency director. The level of discretion was again divided into three categories: narrow, limited, and broad. Under narrow discretion, consultation is mandatory. An example of this is Executive Order 11974 where, before designating or adjusting the original designation of articles eligible for consideration under the Generalized System of Preferences, the president is required to take into account information and advice from the International Trade Commission; the Departments of Agriculture, Commerce, Defense, Interior, Labor, State, and the Treasury; and the Special Representative for Trade Negotiations; and to hold public hearings. ¹³ Under limited discretion, Congress recommended (but did not require) the president to seek advice from a commission, agency head, or department secretary prior to taking action. If no controls were imposed, then the president was coded as having broad discretion. An example is Executive Order 12005, ¹⁴ which allowed the president to redefine an existing item on the tariff schedules of the United States but did not require or recommend seeking advice or information prior to taking action. ¹⁵

In most of the orders presidents issue, the president acts with broad discretion delegated to him by Congress. Specifically, the president’s order issued has broad discretion (where the president has more control over the policy without limitations from Congress) in 73% (2,294) of the substantive cases and 88% (2,752) in the ex-ante cases. Presidents issue orders under the most limited type of discretion when it comes to substantive authority, where they have limited discretion (the requirement to follow a guideline or numerical limit) in 7% (224) of the total cases and limited discretion under ex-ante discretion in 9% (268) of the cases. These trends make for a strong test of the above expectations as presidents are overwhelmingly likely to have the authority to act with broad discretion but there is also variation in how they act in terms of the type of discretion relied upon. If presidents are less likely to act with broad authority, it would be an interesting and counterintuitive finding and would countermand the trend of presidents acting generally with greater discretion, given the range of options they may use to justify their orders. Significant orders present a slightly different story. Presidents act with broad discretion in 88% (172) of the significant orders and 95% (186) of the significant ex-ante orders. With both total and significant orders, given the amount of discretion presidents tend to use, there is clearly less variance to explain. This provides a conservative test case because if we find variation we will have greater certainty that the deviations are meaningful.

Several explanatory variables were included to allow us to test our assertions pertaining to whether or not presidents are issuing unilateral orders under broad discretion with respect to disagreement between or within the branches. We measure interbranch distance (distances between the president and each median in each Congressional chamber or the median of the Court) using common space measures from M. A. Bailey and Maltzman (2008) for interbranch distance for both Congress and the Court. In each instance, we operationalized this by taking the absolute value of the difference between the president and the average of the medians of the majority party members in each chamber. ¹⁶

Several control variables are also included in the models. First, second half is coded as the second half of a president’s term (either first or second term; “Second Half”). Second, second term is coded as any second term for a president if he had one (“Second Term”). Third, election year is coded as any full year in which there is a presidential election (“Presidential Election”). Mayer (1999, 2001) argues that these temporal moments may increase the number of orders, making them an important control. Fourth, we include a dummy variable for moments where a president of a new party enters office (“New President”), starting with the first Congress over which they governed (coded “1” if yes and “0” if no). This variable was included by Howell (2003) as a control because new presidents often overturn prior orders increasing the number of total orders in that year. If the United States is involved in an armed conflict, we include a dichotomous indicator for war (“War”). Young (2013) finds that conflicts like war increase the number of orders, making this an important control. The war variable was coded “1” for the following conflicts: World War II (1941-1945), Korean War (1950-1953), Vietnam War (1964-1973), Persian Gulf War (1991-1992), Afghanistan War (2001-2009), and Iraq War (2003-2009). Finally, we include a variable on the distance between the year the president issued the order and the year the statute took effect (“Statute Date Length”). President may rely more heavily on recently passed statues (Belco & Rottinghaus, 2017) necessitating some indication of the role played by the longevity of the statue in question.

The relevant portions of the statues and changes were compiled from the United States Code and the U.S. Statutes at Large to ensure that the law was captured as it was for the correct time period. If the statute was amended, the most recently amended date was used. The summary statistics of these variables are listed in Table 2a in the online appendix.

When Do Presidents Act With Broad Discretion?

Table 1 specifies several estimated models to test the above expectations. The categories measuring the dependent variable (level of discretion) are ordinal measures as identified in the text above for each of the two types of authority (substantive and ex-ante). ¹⁷ We use several ordered logit regression models, a generalized linear model for a binomial regression, which allows us (a) to utilize a categorized dependent variable for our dependent variable and (b) to determine the probabilistic effect of the key independent variables on the dependent variable (Long, 1997; Agresti, 2002). ¹⁸ The unit of analysis is the individual order. The standard errors are clustered by Congressional term to avoid artificial creation of variation. This design does not explain the use of an order with greater discretion or not, rather it identifies when presidents issue an order, the amount of discretion used.

OPEN IN VIEWER

Table 1.

Ordered Logit Models for Use of Discretion (Bailey & Maltzman, 2008; Common Space Scores).

	Substantive		Ex-ante
	All orders	Significant	All orders	Significant
Supreme Court Distance × Chamber Median	−0.67*** (0.181)	−1.86** (0.828)	−1.13*** (0.228)	−2.74*** (1.13)
Supreme Court distance	3.13*** (0.556)	9.68*** (3.20)	3.28*** (0.540)	6.70*** (2.36)
Distance between Congress median and president	0.017 (0.303)	0.025 (1.22)	0.969*** (0.360)	1.65* (1.25)
Second term	0.067 (0.284)	0.128 (0.747)	0.876** (0.404)	1.25 (1.35)
Second half	0.477** (0.256)	0.478 (0.623)	0.656*** (0.251)	3.99* (2.30)
Presidential election	0.064 (0.192)	0.802 (1.76)	−0.081 (0.186)	−3.86 (3.10)
War	−0.160 (0.317)	1.60 (1.13)	−0.046 (0.341)	1.23* (0.908)
New president	0.063 (0.368)	0.882 (0.901)	−0.066 (0.300)	0.711 (0.931)
Statute date length	−0.001 (0.005)	−0.032** (0.015)	−0.011 (0.006)	−0.037** (0.020)
n	1,412	257	1,412	257
Pseudo-R2	.108	.305	.087	.367
Log likelihood	−907.3	−43.7	−523.7	−18.4
Likelihood ratio χ2	223.1***	129.8***	167.9***	73.7***
Intercept 1	−1.21	2.05	0.187	1.65
Intercept 2	0.598	4.50	0.493	—

Note. Dependent variable: Issuance of an order under different types of discretion, 1951-2009. Standard errors clustered by Congress. Model is ordered logit model. Standard errors in parentheses.

*

p < .10. **p < .05. ***p < .01.

To summarize the expectations from above, we assume that issuing an order with broad authority, even if properly used, risks Congressional reaction or court review. Given this, we expect that presidents are more likely to issue an executive order with broad discretion when there is greater interbranch ideological distance for either a chamber of Congress or the Court; yet, presidents should issue orders using more narrow (less broad) discretion when there is greater interacted interbranch ideological distance for either a chamber of Congress and the Court. Thus, the outcome is not measured by when presidents issue orders with more discretion but rather the degree of discretion used when the order is issued. ¹⁹ Presidents do not stop issuing orders, rather they issue orders using less broad (or more narrow) discretionary authority as they can choose from a menu of options to justify their orders. When the Court and Congress are jointly distant, the probability that Congress will challenge the president, and the challenge will be upheld, increases. In these cases, the president is more likely to issue a narrow justification, even if modestly, in the order that is less likely to be viewed negatively by Congress and more likely to withstand justification by the court. Again, the nature of the issuance of the orders does not allow us to model when presidents do not act, only when they do. Yet, this provides a perspective on how the actions they take is affected by the discretion they claim. This is a higher standard as presidents are more likely to have and cite broad discretion generally—Any politically calculated deviations from broad citations of authority in orders would be substantively significant.

In Table 1, we examine cases where the president faces different political conditions with respect to the other branches for each of the two types of discretion. ²⁰ The results generally confirm expectations from Hypotheses 1 and 2—When presidents are ideologically distant from either the median member of Congress individually (ex-ante, columns 3 and 4) or the Court (substantive, columns 1 and 2; ex-ante, columns 3 and 4), they are more likely to issue orders using broad discretion. ²¹ Specifically, the greater the ideological distance between the median member of the majority party in each chamber or the median member of the Supreme Court, the more likely the president is to issue an order using a statute with broad discretion as signaled by a positive and significant coefficient. As expected, presidents believe they can overcome disagreements with a single branch when exercising greater discretion. The effect is restricted to ex-ante delegation in that there is no statistically significant effect for the distance between Congress and the president for all or significant substantive orders. As more ex-ante orders are issued with broad authority than substantive orders, the influence of closer ideological proximity is enhanced as presidents feel more comfortable issuing orders with broad authority when Congress is ideologically similar to the president.

However, the president’s strategy is altered when both the Court and Congress are jointly farther away from the president’s ideology. ²² In Table 1, the interacted term for the distance between the Court median and the average Congressional median displays a consistent negative effect for all orders and both types of delegation, suggesting that presidents are less likely to issue an order with broad discretion when the interacted difference between the other branches and the executive is greater, consistent with Hypothesis 3. ²³ Given the structure of the coding, which captures the level of discretion used in each order rather than the total discretion allowable, it is not the case that presidents issue fewer orders totally, rather they moderate the discretion they employ when they issue the orders. Put differently, as the distance between the president and the Court increases, conditional on the distance between the president and Congress, presidents use broad justifications less frequently and narrow ones more frequently.

The political synergy between the branches and potential disharmony with the executive’s positions makes the president more fearful of rejection of his preferred policy position and causes him to moderate his use of discretion in the orders. The implication is that substantive and ex-ante Congressional limitations are a restraint on the president’s ability to act with broad authority but only when Congress may object to the employment of the discretion and Court is more likely to overturn the president’s order. Thus, even though Congress gives the president broad authority, he is still not always willing or able to exercise it. Figure 1 graphically summarizes the predicted probabilities from Table 1 for substantive discretion and Figure 2 does the same for significant orders. ²⁴ In each case, presidents were substantially less likely to act with broad discretion for both all and significant orders as the joint ideological distance between the chamber and the Supreme Court and the president increased. Similarly, significant orders are more strongly affected by the joint ideological distance and the probability of issuing an order with broad discretion drops more sharply for these orders for both substantive and ex-ante discretion. The change in probability of using an order with broad discretion is sizable even for small ideological differences. That is, although the joint Congress–Court ideological distance is usually modest (below 2), signaling general agreement between the branches, even small differences produce about a 1% to 15% reduction in the probability of a president issuing an order with broad discretion. Orders using substantive discretion and significant unilateral orders in general are more likely to be affected by ideological differences. Theoretically, as the ideological distance grows, Figures 1 and 2 demonstrate that the probabilities of using broad discretion would continue to fall.

OPEN IN VIEWER

Figure 1.

Predicted probabilities for substantive discretion.

OPEN IN VIEWER

Figure 2.

Predicted probabilities for ex-ante discretion.

The findings support recent interpretations of when presidents will act unilaterally by providing evidence that emphasizes that presidents will use unilateral authority less when they fear Congress and the Court is more likely to challenge the president’s use of the order (Howell, 2005). These findings show that presidents are constrained in using delegated authority by the possibility of legislative and judicial dissent, even when their authority to act is broad. For instance, in the case of ex-ante discretion, when Congress provides for a dual system of delegation, an executive agency is consulted and the executive is the ultimate decision maker. The president may use this information asymmetry to his advantage (Katyal, 2006), especially when presidents have broad authority to act. The president may choose to cite a statute that requires agency oversight of a policy issue based upon the sensitivity of possible restraint from other branches to moderate his use of executive authority and appear to be more compliant. In this example, when the potential for interbranch disagreement is greater, the president may decide to peremptorily limit their own executive power by requiring an agency consult (narrow discretion for the president) before policy action is taken.

Conclusion

The separation of powers as outlined by the Constitution implies more power sharing than power separating (Fisher, 1989). The results in this article demonstrate that the prospect of joint interbranch disagreement influences the level of delegated authority presidents use when issuing unilateral orders. Our findings have shown that, while presidents do act first and alone and cause the other players in the system to react (Howell, 2003), other actors significantly shape the process of when and under what conditions presidents exercise discretion in employing unilateral orders. Congress can be viewed as a first mover in this case, instead of the president. These findings not only supplement our understanding about the nature of presidential unilateral orders and the governing system but also confirm that presidents frequently act unchecked with broad authority. Although presidents may appear to act with unilateral dispatches over the other branches, the reality is often more detailed than can be explained by the mere presence of a unilateral order. As Kiewiet and McCubbins (1991) note, “principals can be highly successful in using delegation to pursue their interests, yet appear passive and ineffectual, relative to their agents” (p. 170). Presidents and their lawmaking partners in Congress and the Courts often agree on a delegation of power, the appropriate balance of the use of this power, and the duration of that authority in a system that separates and shares powers (Dickinson & Gubb, 2016; Ouyang & Waterman, 2015; Rottinghaus, 2015).

In this vein, our findings make two contributions to our knowledge regarding power sharing between the president, Congress, and the Court. The primary contribution of this research is to identify the degree and type of unilateral activity presidents engage in considering key interbranch checks. Although several scholars of unilateral powers describe the restraints of delegated authority and the restrictive element of adherence to legislative statute (Mayer, 2009; Moe & Howell, 1999), no study has used these limitations as a way to describe the exercise of this critical authority. Contrary to the portrayal of Congress as a reactive agent (Black, Madonna, Owens, & Lynch, 2007), the picture presented here suggests the use of unilateral orders as the outcome of political bargaining (Dickinson, 2008). Congressional response to executive orders is found to be minimal perhaps because no reaction is needed when Congress retains some control over the process before the president even issues an order and where presidents moderate their use of power voluntarily. Ultimately, the factors that influence the president are political, but they are also structural. Presidents are fearful of a negative reaction from Congress and the Court even when they may possess greater discretion. Constrained or unconstrained, however, presidents still do not hesitate to act with broad, substantive and ex-ante discretion delegated to them by Congress, especially when there is ideological friction between the president and the Congress or the president and the Court (but not both).

The second contribution of this research is to expand on our knowledge of how the system of checks and balances operates on the president’s exercise of discretion. The bounds of discretion on the executive are generally defined by the checks placed upon it by the other two branches of government. Presidents do not always have significant limits placed on the exercise of discretion nor are Congress or the Court without capacity to respond to presidential independence. Indeed presidents are granted a great degree of broad discretion for most orders. In the absence of statutory constraints, Congress can enact new legislation limiting presidential power, and with procedural constraints, it can curb executive action with a legislative override. The reality of the president’s exercise of discretion reveals that the use of power is restrained by joint ideological institutional disagreement even if the president is more likely to act with broad authority when single interbranch disagreement is greater. This finding helps to address the debate about when presidents are likely to issue unilateral orders—There is support for both an assertive and restrained president with respect to the amount of discretion delegated and institutional friction.