Printer Friendly

After the orders: presidential memoranda and unilateral action.

The "Mexico City Policy," which prohibited U.S. foreign aid to organizations that funded abortions, has been the subject of an executive tug-of-war since established by the Reagan administration. Bill Clinton repealed the policy in one of his first actions as president (Clinton 1993). In 2001, George W. Bush reinstated the policy, writing that it was his "conviction that taxpayer funds [...] should not be given to foreign nongovernmental organizations that perform abortions or actively promote abortion" (Bush 2001a, 17303). President Barack Obama eliminated it for the second time in his first days in office (Obama 2009). These actions constitute important examples of what might be termed "unilateral action." (1) However, none of them were taken by the most familiar means of such action--the executive order. In each case, policy change occurred by way of presidential memoranda', a class of presidential actions that contains orders to administrators and is not subject to the statutory reporting requirements of executive orders and proclamations.

In response to a relatively constrained view of presidents--forced to "bargain" their way to policy accomplishment--an important vein of presidency scholarship has shifted focus to the president's formal and informal instruments of unilateral action. This has led to systematic analyses of the tools available to the president: executive orders; proclamations; and, most recently, signing statements. The aim of this article is to consider this additional tool--presidential memoranda--and to suggest what the study of unilateral action, more broadly, could gain by including them. To preview those conclusions, the payoff of including presidential memoranda is that they reveal additional observations of unilateral action and provide initial analytical leverage for understanding what induces presidents to use one tool over another.

I argue that these memoranda contain important policy content, and that in recent years, they have been used interchangeably with executive orders. Memoranda perform a variety of functions, however, like executive orders, they are essentially orders to administrators. Researchers typically reference executive orders to highlight the degree to which the president is capable of making policy independent of other institutional actors. This extends to a subset of presidential memoranda, as it does for a subset of executive orders. Presidents have used memoranda to set policy, sometimes acting independent of authorizing legislation. In light of this, it is significant that over the last 50 years, published memoranda have surged as executive order issuance has declined.

Analyzing memoranda issued between 1946 and 2013,1 find evidence that memoranda are increasingly significant, measurable outputs of executive action. Presidents have increasingly used memoranda rather than executive orders to effect similar ends. (2) Moreover, memoranda and executive orders both appear to be indications of presidential unilateralism. Taking into account variables, which others have shown to effect the frequency of executive orders from year to year, memoranda are positively related to their more well-known counterpart.

The observed relationships highlight the malleability of these administrative tools--which often have few formal, legal standards. Far from being fixed, presidents largely determine the definition of these instruments and how they ought to be used. More specifically, the increased frequency and importance of presidential memoranda suggests changing incentive structures have contributed to a corresponding reduction in executive orders. Presidential memoranda may be a less politically costly means of action as public and elite awareness of executive orders have increased. Contemporary news coverage of executive orders is ubiquitous, such that the tool immediately evokes potentially damaging questions of "imperial overreach." (3) Recent events during the first and second terms of the Obama administration indicate that the shift between executive orders and presidential memoranda may mirror the move from signing statements to "statements of administrative policy"--which, as Crouch, Rozell, and Sollenberger (2013) note, perform similar functions without the negative publicity associated with signing statements since the end of the second Bush administration.

The article proceeds in a few basic steps. First, I consider presidential memoranda in the context of unilateral action theory, arguing they constitute additional observations of the theory's outcome of interest. Second, I lay out a few basic expectations about memoranda and executive orders designed to test the assumption that they are used in systematically similar political contexts. Third, I provide a cursory overview of how presidents since Jimmy Carter have used memoranda, ending with the most recent years of the Obama presidency--when published presidential memoranda have eclipsed executive orders in frequency. Fourth, I run a systematic analysis of memoranda published since 1946, finding evidence that memoranda and orders vary predictably from year to year. I then consider possible implications for future studies of presidential documents and unilateral action, and argue that the rise in memoranda is the result of changing political incentives associated with executive orders.

Unilateral Action Theory

As Howell and Lewis (2002) note, strategic employment of unilateral actions in the president's toolkit is one of the defining features of the modern presidency. Systematic consideration of these tools began in earnest with executive orders, a literature that has now expanded considerably (Fine and Warber 2012; Glieber and Shull 1992; Howell 2003; Mayer 1999, 2001; Mayer and Price 2002; Rudalevige 2012; Warber 2006). Mayer and others in this area sought to reassess the former "paradigm of the presidential studies literature that [held] that presidents have limited capacity to act unilaterally"(1999, 445). More recently, scholars have branched out into signing statements (Crouch, Rozell, and Sollenberger 2013; Kelly and Marshall 2007, 2008; Korzi 2011; Ostrander and Sievert 2013) and proclamations (Bailey and Rottinghaus 2013; Rottinghaus and Maier 2007). While these studies do represent an important step toward defining the president's means of acting beyond informal bargaining (which is difficult to observe), the research program as a whole has yet to systematically analyze memoranda use across time.

Most surveys of the tools of presidential action leave out memoranda. Of those that do not, typically, they are briefly mentioned. Mayer (1999) writes,

   The major classes of presidential policy instruments are executive
   orders, proclamations, memoranda, administrative directives,
   findings and determinations, and regulations. Of these, executive
   orders combine the highest levels of substance, discretion, and
   direct presidential involvement. (35)

While this conclusion is now standard and somewhat warranted, memoranda remain an important component of presidential action. Since then, memoranda have been treated most comprehensively by Cooper (2001, 2002), who has argued they are sometimes part of "shell games" intended to deceive organized interests, wherein presidents issue highly visible executive orders, only to then issue memoranda with very different policy implications. Cooper notes a variety of other uses: to generate positive publicity, to initiate policy change, and to manage emergencies. Most pertinent to this analysis, however, is Cooper's observation that "as with other tools of presidential direct action, the precise definition of the presidential memorandum is unclear and evolving" (2002, 83). This is one explanation as to why these memos have yet to be fully integrated into the ongoing scholarly discussion of unilateral power. Memoranda pose significant data collection challenges and are subject to considerable definitional variance.

However, the fact that presidential memoranda should be included in that discussion is plainly apparent when one considers the basic points of theories of unilateral action. Moe and Howell (1999) provide the clearest explanation of those points, writing,

   Presidents have incentives to expand their institutional power, and
   they operate within a formal governance structure whose pervasive
   ambiguities--combined with advantages inherent in the executive
   nature of the presidential job--give them countless opportunities
   to move unilaterally into new territory, claim new powers, and make
   policy on their own authority. (871)

Presidents have institutional advantages vis-a-vis Congress that they can and do exploit to get what they want. (4) Howell's (2003) work helped reframe the contemporary consideration of presidential power by shifting attention from negotiation and persuasion (Neustadt 1960) to the positive actions presidents take with some degree of independence. (5) The literature focusing on presidential documents, then, is concerned with measuring and counting "outputs" as indications of the degree of unilateralism in each administration. Presidential memoranda are a vital part of these outputs. The substantive significance of this tool has increased, just as its issuance has become more routine. The most illustrative way to investigate this is to consider important uses of presidential memoranda across time and administrations. However, in order to demonstrate systematically that memoranda and executive orders have been used in similar political contexts, I also include a statistical consideration of presidential memoranda issued from 1946 to 2013. For that analysis, I have two basic expectations. First, that executive orders and presidential memoranda will be positively correlated. Increased frequency of one will lead to increased frequency of the other. As the proceeding cases illustrate, memoranda and orders are often used in conjunction to effect policy change. Moreover, I take these two tools to be indicators of the same underlying phenomenon, so that the presence of one should increase the likelihood of the other. That is, if executive orders and memoranda are both outputs of unilateral action, then we should expect them to be positively correlated. It is necessary to pause to distinguish this expectation from the developmental trend I later identify. Scholars have identified a decline in executive order use in the postwar era. Here, I find a rise in memoranda use over the same period. This trend is important, in that it may indicate a transition in the types of tools presidents use to act alone. But in order to show that yearly memoranda use exhibits patterns others have found in executive orders--a substantively different question--I must set aside that trend. This involves a statistical adjustment conventionally applied to trending data, which I discuss later.

Second, I expect that memoranda will vary predictably based upon familiar covariates of executive orders. If memoranda perform similar functions and are both indicators of unilateralism, then I should observe presidents issuing them in similar political contexts. Among other circumstances, presidents have been shown to issue more executive orders if they are Democrats, under unified government, and in the last two years of their presidency. [H.sub.1] follows from my first expectation:

[H.sub.1]: Higher counts of executive orders will be associated with higher counts of memoranda.

In addition to being positively correlated, I expect that Democrats will issue more memoranda than Republicans. Past work has found consistent evidence that Democrats tend to issue more executive orders. I expect this trend will translate, since memoranda, like orders, may be indications of more active administration, relating to the demands of interests associated with the Democratic Party (Deering and Maltzman 1999; Fine and Warber 2012; Mayer 2001).

[H.sub.2]: Democratic presidents will issue more memoranda.

Mayer (2009, 2010) highlights a consistent empirical puzzle in the executive order literature: that their use tends to decline during divided government, contrary to the predictions of unilateral action theory. If unilateral action really is about circumventing Congress, then preference divergence between the two institutions should induce such action. More recently, Fine and Warber (2012) have addressed this puzzle by distinguishing between symbolic, routine, and major policy orders. They find the more intuitive effect that presidents issue more policy orders during divided government. Fine and Warber make internal distinctions between executive orders that I do not apply to memoranda, so it is quite possible that memoranda and divided government will exhibit their counterintuitive relationship. However, I retain the expectation which follows from unilateral action theory:

[H.sub.3]: Presidents will issue more memoranda during divided government.

I expect that presidents will issue more memoranda in the latter half of their terms, in part because of speculative seasonality related to unilateral action. If presidents bare costs for circumventing Congress, then it is possible presidents will only resort to such action when subject to end-of-term time constraints, and long after the political capital of their first one hundred days has diminished. Additionally, this variable will pick up, with some regularity, a campaign effect--whereby presidents will issue more memoranda in order to reap electoral gains associated with policy change.

[H.sub.4]: Presidents will issue more memoranda in the final two years of their term.

No consensus exists on the role of public opinion in unilateral action. Studies have exhibited mixed findings relating to presidential approval ratings and executive orders (Deering and Maltzman 1999; Fine and Warber 2012; Krause and Cohen 1997; Mayer 2001). Here, I expect that lower approval ratings will be associated with more memoranda. Policy change by administrative fiat often identifies issue areas important to specific constituencies--so that unilateral action becomes a means of shoring up electoral support. (6) Thus, I expect presidents looking to respond to waning popularity will become more energetic executives.

[H.sub.5]: Presidents will issue more memoranda when their approval rating decreases.

I also expect that--given that much of the policies set by presidential memoranda are only as durable as the administration that issued them--I will observe more memoranda when the president's partisan identification changes. Presidents riding the wave of an electoral "mandate" to repudiate the opposition may seek to stake their positions and differentiate themselves from previous administrations (Mayer 1999). This article began with a case in point, wherein each successive president reversed the policy on foreign aid for organizations that perform abortions.

[H.sub.6]: Presidents will issue more memoranda in the first year of their term if the previous president had a different partisan identification.

The Development of Presidential Memoranda

Presidential Memoranda have been used to perform a myriad of functions, and unlike executive orders, no formal (or legal) requirements have been attached to their use. Memoranda have been used to delegate authority to subordinates, make determinations relating to existing statutes, create task forces, designate acting heads of agencies, request agencies review decisions, order agencies to make policy in a given area, and change administrative policy unilaterally. Cooper terms presidential memoranda, "executive orders by another name and yet unique"(2002, 80). Memoranda do contain important policy content pertinent to the study of unilateral action, such that it is often difficult to discern the difference between memoranda and executive orders, were one to read them side by side. By way of introduction and comparison, Figure 1 plots the number of memoranda and executive orders over time. These memoranda were collected using compilation tables of presidential documents, as well as the digitized Federal Register. To my knowledge, it is the first data set of presidential memoranda to include the entire postwar period. An additional caveat is that this initial graph forgoes any attempt to distinguish between "significant" or "important" orders and memos--it merely reports totals from year to year. I detail how these data were collected and discuss my decision to retain the aggregate totals in the next section. Still, if we briefly entertain the assumption that orders and memos contain roughly equivalent proportions of impactful issuances, then the trend lines provide an initial confirmation that the unilateral action literature may be enriched by including presidential memoranda. Noticeably, the executive order count has trended downward for the past 50 years as memoranda have increased in frequency. Experts in this area have noted the decline of executive orders in recent years, together with the rise of alternative tools (Rudalevige 2014). It may be that the lesser-known presidential memorandum has partially supplanted the executive order. That is, unilateral action is not declining; presidents have merely shifted the means by which such action is taken.

Important cases of memoranda use up to the Clinton and early Bush administrations have been dealt with in thorough detail by Cooper (2002). Among the cases that stand out prior to the Clinton administration are Carter's series of administrative review memoranda (7), Ronald Reagan's hiring freeze and regulation postponement (8), and George H. W. Bush's implementation of de facto statehood for Puerto Rico. (9) Clinton's memoranda addressed a variety of issues, many of them with explicit connection to the Democratic platform: "privately funded abortions at military hospitals," Title X, federal funding of fetal transplant research, and the importation of the RU-486 "abortion pill" known as Mifepristone. Famously, current Supreme Court Justice Elena Kagan (2001) defended the Clinton administration's routine and effective exercise of administrative discretion designed to work toward policy goals--often with memoranda as the tool of choice. Other memoranda were issued in the areas of gun control (August 11, 1993), fair housing (January 1, 1994), immigration (February 2, 1995), domestic violence (October 2, 1995), and air pollution (July 16, 1997). Additionally, Clinton issued memoranda similar to Carter's internal reviews, pertaining to rule making (September 30, 1993), management reform (October 1, 1993), and "streamlining the bureaucracy" (Clinton 1993). Though the goal of this study is not to provide in-depth case studies of memoranda use, it is important to provide a few contemporary examples during the Obama administration that substantiate this trend.

The scope and policy content of the memoranda issued by the Obama administration is not new, but the frequency of their issuance is unprecedented. To give rough idea of this scale, consider the fact that published memoranda began eclipsing executive orders in frequency for the first time during the Obama presidency. (10) Like previous administrations, Obama used memoranda to address current political debates, overturn previous president's policies, and signal to constituents his administration intended to follow through on campaign promises. Reminiscent of Reagan's hiring freeze, the Obama administration issued a pay freeze the first day in office, along with reform memoranda relating to government transparency. Memoranda in this vein promised to expedite review processes for federal work projects, provide flexibility in the implementation of environmental regulations, and curb discretionary bonuses to political appointees. Using memoranda, the administration announced task forces on childhood obesity, the middle class, federal contracting with small businesses, the space industry, and job training. (11)

Obama's administration consistently issued policy-related memoranda throughout his first term, especially during 2009 and 2010. During 2011 and 2012, the frequency of these memoranda went down, but the political context essential to the strategic model of unilateral action was put front and center by the administration's "We Can't Wait" campaign--a set of 45, purposely publicized executive actions. Six of these began with a presidential memorandum. Other actions related to six executive orders. In a press conference on October 26, 2011, Obama stated its purpose quite succinctly: "We can't wait for Congress to do its job. So where they won't act, I will" (Wolf 2011). In 2013, following a series of tragic mass shootings, the Obama administration issued two memoranda: one that required the Secretary of Health and Human Services to fund research to investigate the mental health causes of gun violence, another that amended the Bureau of Alcohol, Tobacco, and Firearm's process for tracing firearms. These actions seem to comport with the basic prediction of unilateral action theory, since that same year, Congress did not change policy relating to federal background checks.

These cases provide an initial confirmation of the idea the presidential memoranda have the potential to result in policy change both alongside and independent of executive orders. The memoranda are neither wholly clerical nor merely hortatory. In a basic sense, they provide orders to administrators to carry out presidential policy goals. Though the cases give some indication of these points, it is equally important to determine whether memoranda use varies from year to year and administration to administration in predictable ways. In particular, if they are indications of the same underlying phenomenon, then they ought to be wedded to similar covariates.

Data and Estimation

The memoranda analyzed herein were collected using the compilation tables of the Code of Federal Regulations and the digitized Federal Register, (12) In the Code, presidential memoranda are listed under "Presidential Documents Other than Proclamations and Executive Orders" within Title 3. (13) At the National Archives, these memoranda are indexed under "government documents having general legal effect." (14) Here, it is essential to note the standards of inclusion used. Memoranda (and several other types of actions) pose a difficult research challenge because of the historical fluidity of what presidents have chosen to call them. As recent as 1978, presidents were issuing them under their former name, "presidential letters." (15) From 1959 to 1978, each president issued both "letters" and "memoranda"--though there is no apparent legal or substantive difference between them. (16) Additionally, another class of presidential actions, "determinations," are inconsistently commingled with memoranda. Cooper (2002) identifies these determinations as "standard foreign policy" tools--a subset within presidential memoranda. In the majority of cases, determinations were labeled. In the digitized Federal Register, these have been categorized as a separate tool--as opposed to a subset within memoranda. I treat them as such--in part, because determinations (even if labeled as memoranda) have both a consistent substantive content and (more recently) a separate codification process. (17) Finally, in some cases, the title of the document within each table had no label at all. In each case, I consulted relevant Government Printing Office and Federal Register publications, where the "memoranda" label typically appeared in a subtitle. To recap, I included a document listing if

1. The document was labeled either "letter" or "memorandum."

2. The document did not contain a determination.

3. The document was not labeled but was referred to as a memorandum in text.

This collection process yielded a total of 693 presidential memoranda published in the Federal Register from 1946 to 2013. Though this total is dwarfed by executive orders (over 4,000), more than half of those memoranda have been issued since 1995.

Before proceeding, it is important to note that the means of inclusion I have outlined leaves out many more memoranda, which the presidents have not directed to be published in the Federal Register. The decision to publish a memorandum is left entirely to the president, unless the action contained within it requires publication by law. This poses a considerable data collection challenge, since the remainder exist within the unindexed Weekly Compilation of Presidential Documents, (18) The question, then, is how the subsample I have described restricts the conclusions one may draw about the universe of presidential memoranda. The most pertinent feature of the selection process is that the discretion over whether to publish lies with the White House. The memoranda analyzed are thus, the subsample of memoranda the president has chosen to publish. Revisiting the cases I highlight in the previous section, this is not particularly surprising. These memoranda often have clear connections to policy areas and their relevant constituencies. (19) Since the proceeding analysis sets these memoranda alongside executive orders, I argue the selection process does not pose substantial problems. Presidents were legally obligated to publish every executive order in the timespan I consider. (20) Thus, I compare two classes of documents that were published in the same venue. These actions were informed by the knowledge that they would be made public at some initial level. Whatever additional publicity these actions receive is, as I argue, partly a function of increased salience of executive orders among the media and the public. Finally, while the analysis of memoranda that were intentionally excluded from the Federal Register could pose its own unique set of research questions, they are beyond what I consider here. (21)

Landmark research on unilateral action and executive orders has developed a variety of systematic methods for determining what presidential actions should be deemed "significant." Mayer (1999) began by omitting executive orders dealing with mandatory retirement and public lands. Mayer and Price (2002) developed a more extensive process, modeled after Mayhew's (1991) "sweep method," and designed to pick up orders referenced by Congress, scholars, presidents, or litigation. Howell (2003) employed a similar method, with orders referenced in the Congressional Record or multiple federal court opinions meriting inclusion. These significance tests may be inappropriate for presidential memoranda. Sweep One--for contemporary sources--may exclude many important memoranda, which are relatively unknown to news media and the public. As Cooper (2002) notes, media sources falsely reported some of Clinton's memoranda as executive orders--a trend that has continued into the Obama presidency. (22) Retrospective accounts may also be problematic, since a substantial proportion of published memos were issued within the last 10 years. Since my data set is already a subsample of all memoranda (those that were published in the Federal Register), I have opted to include all memoranda initially. Rudalevige (2012) made a similar decision in his recent investigation of the origins of executive orders, citing the possibility that seemingly mundane functions like exemptions from retirement could have significant policy effects. Additional work designed to make internal distinctions between executive actions may be a next step in the study of presidential memoranda. As Chiou and Rothenberg (2014) note, the connection between presidential action and presidential power is often difficult to parse out. I have provided a few substantive examples in which that has clearly been the case, but ultimately, a more complete analysis of the connection between presidential actions and policy outcomes is beyond the scope of this article. Given the systemic problems with applying existing "significance tests" to memoranda, a uniform procedure for content analysis may be most appropriate. (23)

Despite the concerns I lay out above, I have provided an additional specification of the dependent variable (the yearly total of memoranda), which eliminates three types of memoranda that may be clerical in nature, depending upon one's definition of the term. This total eliminates delegations, certifications, and designations. In the results section, I refer to these as "alternatively specified" (alt.) variables. Some cases of delegation include seemingly trivial matters, in which the president gives a subordinate authority to report to Congress when a statute explicitly demands the president. Certifications also directly relate to statutory requirements imposed by Congress. The president designates acting subordinates in the case of vacancies. It is important to note that none of the relatively few cases of designations in memoranda could be considered recess appointments. In order to keep this specification comparable, I performed the same process with executive orders in the model that includes it. I include these specifications--with each type omitted--as a robustness check against the bare totals.

I acquired all executive orders issued from 1946 to 2011 through the Policy Agendas Project from the University of Texas at Austin. More recent years (2012-2013) were, again, provided by the federal Register. Yearly presidential approval ratings were obtained from the Roper Center's Public Opinion Archives. (24) I also included indicator variables for divided government, each president, the first two years of each term, and the first year of a term in which the president's partisan identification changed. Each is designed to test the expectations I have laid out in the previous sections. Additionally, presidential fixed-effects isolate the unobserved characteristics that may be unique to presidential administrations--even though those characteristics remain largely speculative and unexplained by the analysis.

The linear equation below is a part of a negative binomial model. However, employing a standard event-count model is clearly insufficient in this case. As Figure 1 illustrates, both memoranda and orders contain nearly monotonie trends over the period in question. That is, any variable with some time trend could be modeled as a statistically significant covariate of the count of memoranda. In order to purge both variables of serial dependence, I have included lagged-memoranda and a trend-corrected version of the executive orders count based upon the Box and Jenkins' (1976) procedure for time-series data. Accordingly, the executive orders variables are the residuals of the executive order count, modeled as an AR(2) process. In this way, the results investigate patterns of memoranda and executive order use beyond the observation that follows from Figure 1: that memoranda appear to be replacing executive orders. That is, I control for the developmental trend between these tools in order to investigate the year-to-year political contexts in which they are used.


To recap, I expect a positive relationship between yearly presidential memoranda and executive orders. I expect Democratic presidents and those under unified government to issue more memoranda. Additionally, more memoranda should be issued in the last years of a presidential term, when the president's popularity wanes, and when there is a change in the partisan control of the White House.


The available data provide preliminary support both for the complementary effect between--and rough equivalence of--presidential memoranda and executive orders. Overall, memoranda and executive orders follow similar patterns in terms of yearly counts. The returned maximum likelihood estimates are reported in Table 2, while more readily interpretable marginal effects estimates are reported in Table 1. When memoranda and orders containing delegations, designations, and certifications are removed, most of the results actually strengthen. As expected, after purging executive orders and memoranda of their time-related disturbance, the two tools vary together--with higher counts of one associated with higher counts of the other ([H.sub.1]). This result is similar in magnitude for the alternative specifications of both variables. In terms of the documents themselves--when one controls for the developmental convergence of the two tools and other factors, a president who issues an additional 20 executive orders is likely to issue three more memoranda under unified government. (25) In conjunction with the developmental trend apparent from Figure 1, these tools seem to be related to the same kinds of political contexts.

Prior to considering the other results, it is important to pause and provide clarification. To some, it may seem that the positive empirical relationship between orders and memoranda I find belies the apparent relationship in Figure 1. Reasonably, one might ask how could memoranda be simultaneously replacing executive orders and positively correlated with them? By de-trending the data, I remove systemic evidence of the broader developmental trend--the substitution of orders for memoranda--in order to get at the year-to-year patterns of unilateral action. To put it simply, while Figure 1 provides evidence that memoranda may be replacing orders, Tables 1 and 2 suggest they are both indications of unilateralism.

Consistent with executive order use, Democratic presidents seem to issue more memoranda ([H.sub.2]). It is important to note, however, that this ordering is largely an additional illustration of the developmental trend--in that, the Obama administrations' pathbreaking use of memoranda appears to be driving the results. To put this in perspective, the Clinton administration (notable for its use of policy-related memoranda) issued, all else equal, about 16 fewer memoranda per year. Additionally, divided government is associated with substantially fewer memoranda--an effect that nearly doubles when the alternative specification is used ([H.sub.3]). While an additional 20 orders is associated with three more memoranda under unified government, presidents are only likely to issue one more memorandum during divided government. On average, presidents issue five fewer memoranda per year during divided government. This finding is consistent with the persistent empirical puzzle Mayer (2010) highlights--that the number of executive actions actually decreases during divided government. Though this result is contrary to theoretical expectations, it shows memoranda follow an empirical pattern similar to that of executive orders.

The last two years of a president's term are associated with more memoranda, when compared to the first two ([H.sub.4]). Again, this relationship only strengthens when potentially clerical memoranda are removed. All else equal, presidents issue an average of two more memoranda per year in the last two years, compared with the first. Importantly, Mayer (2010) found similar patterns for executive order issuance for both the president's party, divided government, and the year in term. There was no statistically discernible relationship between presidential approval rating and memoranda issuance ([H.sub.5]). This result is not particularly problematic when it comes to the more general relationship I posit between orders and memoranda, because the specific mechanism by which public opinion would affect unilateral action remains unclear. Both popular and unpopular presidents may have incentives to act alone. Additionally, there was no statistically discernible relationship between a change in the president's party and the frequency of memoranda ([H.sub.6]). Cases like the Mexico City Policy appear to be difficult to pick up with any regularity.

Finally, it is important to note the variation in memorandum use between presidents substantially changes when one looks at the alternative specification results in model 2. This point highlights the importance of presidential fixed effects, with the basic explanation that patterns of memoranda (and executive order) use can partly be explained by features unique to individual administrations. For example, the marginal effect of George W. Bush's administration changes--from issuing 16 fewer memoranda per year than Obama under the first specification, to 20 fewer under the alternative--in part, because the administration issued an abnormal number of delegation memoranda. Though broad systemic variations can be discerned based upon political context, there are still unobserved features of particular administrations that lead them to use these tools in idiosyncratic ways.


This article began by arguing that a more comprehensive understanding of unilateral presidential action could not be attained without inclusion of presidential memoranda. To that end, I have shown that memoranda and executive orders have the potential to be equally significant. They both provide instructions to administrators in pursuance of the president's policy goals. Moreover, published memoranda have dramatically increased in the last twenty years. In the most recent administration, they have been used almost interchangeably with executive orders. Key cases of memoranda use, as well as the broader frequency of this tool over the past 50 years indicate that memoranda may be replacing executive orders. However, simultaneously, I have found evidence they may be indicators of an underlying degree of unilateral action. This finding is a first step toward understanding what moves presidents to choose one tool over another. I argue, in the case of orders and memoranda, contemporary movement toward published memoranda is partly a function of increased potential costs associated with executive orders. (26) Whereas executive orders have attained significant media coverage and have drawn litigation, presidential memoranda are frequently overlooked by news outlets and have not been subject to direct legal action.

Moreover, the memoranda-orders dyad may parallel that of signing statements and "statements of administrative policy." Crouch, Rozell, and Sollenberger (2013) suggest the latter may simply be signing statements by another name. Here, the movement toward the more nebulous "statements of administrative policy" may be the result of negative publicity and criticism of the Bush and Obama administrations' use of signing statements (Savage 2006, 2009, 2010). The accumulation of this kind of media attention, together with the potential for legal challenge, places strong incentives on presidents to find new, more innovative and obscure means of acting alone. In this case, the present obscurity of presidential memoranda may allow presidents to claim credit for policy change, while avoiding the charges of "imperial overreach" likely to be levied by critics. In this way, the tools of unilateral action are somewhat endogenous to the political actors who use them. They are not fixed instruments.

The preceding analysis suggests that understanding unilateral presidential action may require some degree of aggregation. That is, the various tools must be grouped, according to the functions they perform and the costs associated with them. This article has provided a preliminary answer to the question of why presidents use memoranda as opposed to executive orders, and a similar answer may apply to signing statements and statements of administrative policy. Though valuable research has considered presidential documents (executive orders, proclamations, and signing statements) in relative isolation, it is clear that each comprises only a part of the president's "toolkit." At the very least, future work ought to consider pairing these tools to get a better indication of the universe of actions--since omitting them might obscure unilateralism where it exists. More broadly, the study of the contemporary presidency will benefit from considering the tools of unilateral action as their institutional cost, substantive effect, and incidence changes over time.


Aberbach, Joel D., and Bert A. Rockman. 1999. "Hard Times for Presidential Leadership? (And How Would We Know?)." Presidential Studies Quarterly 29 (4): 757-77.

Ayres, B. Drummond. 1981. "Civilian Hiring Freeze Called a Success." New York Times, February 10. (accessed January 20, 2014).

Bailey, Jeremy D., and Brandon Rottinghaus. 2013. "The Development of Unilateral Power and the Problem of the Power to Warn: Washington through McKinley." Presidential Studies Quarterly 43 (1): 186-204.

Box, George E. P., and Gwilym M. Jenkins. 1976. Time Series Analysis: Forecasting and Control. San Francisco: Holden-Day.

Bush, George H. W. 1992. "Memorandum on the Commonwealth of Puerto Rico." November 30. (accessed July 18, 2014).

Bush, George W. 2001a. "Restoration of the Mexico City Policy." Federal Register 66: 17303-313.

Chiou, Fang-Yi, and Lawrence S. Rothenberg. 2014. "The Elusive Search for Presidential Power." American Journal of Political Science 58 (3): 653-68.

Clinton, William J. 1993. Memorandum on the Mexico City Policy. Public Papers of the Presidents of the United States, William J. Clinton, 1993, Book 1, January 20 to July 31, 1993? Washington, DC: U.S. Government Printing Office.

Cooper, Phillip J. 2001. "Presidential Memoranda and Executive Orders: Of Patchwork Quilts, Trump Cards, and Shell Games," Presidential Studies Quarterly 31 (1): 126-41.

--. 2002. By Order of the President'. The Use and Abuse of Executive Direct Action. Lawrence: University Press of Kansas.

Crouch, Jeffrey, Mark J. Rozell, and Mitchell A. Sollenberger. 2013. "President Obama's Signing Statements and the Expansion of Executive Power." Presidential Studies Quarterly 43 (4): 883-98.

Deering, Christopher J., and Forrest Maltzman. 1999. "The Politics of Executive Orders: Legislative Constraints on Executive Power." Political Research Quarterly 52 (4): 767-83.

Dodds, Graham. 2013. Take Up Your Pen: Unilateral Presidential Directives in American Politics. Philadelphia: University of Pennsylvania Press.

Fine, Jeffrey A., and Adam L. Warber. 2012. "Circumventing Adversity: Executive Orders and Divided Government." Presidential Studies Quarterly 42 (2): 256-74.

Gleiber, Dennis W, and Steven A. Shull. 1992. "Presidential Influence in the Policymaking Process." Western Political Quarterly 45: 441-67.

Howell, William H. 2003. Power without Persuasion: The Politics of Direct Presidential Action. Princeton, NJ: Princeton University Press.

Howell, William G., and David E. Lewis. 2002. "Agencies by Presidential Design. "Journal of Politics 64 (4): 1095-114.

Kagan, Elena. 2001. "Presidential Administration." Harvard Law Review 114 (8): 2246--385.

Kelley, Christopher, and Bryan W. Marshall. 2007. "Like Father, Like Son: The Presidents' Bush and the Signing Statement." White House Studies 7.2.

--. 2008. "The Last Word: Presidential Power and the Role of Signing Statements," Presidential Studies Quarterly 38 (2): 248-67.

Korzi, Michael J. 2011. "'A Legitimate Function': Reconsidering Presidential Signing Statements." Congress & the Presidency 38 (2): 195-216.

Krause, George A., and Benjamin F. Melusky. 2012. "Concentrated Powers: Unilateral Executive Authority and Fiscal Policymaking in the American States!' Journal of Politics 74 (1): 98-112.

Krause, George A., and David B. Cohen. 1997. "Presidential Use of Executive Orders, 1953-1994." American Politics Quarterly 25 (4): 458-81.

Mayer, Kenneth R. 1999- "Executive Orders and Presidential Power. "Journal of Politics 61 (2): 445-66.

--. 2001. With the Stroke of a Pen: Executive Orders and Presidential Power. Princeton, NJ: Princeton University Press.

--. 2009. "Thoughts on 'The Revolution in Presidential Studies.'" Presidential Studies Quarterly 39 (4): 781-85.

--. 2010. "Going Alone: The Presidential Power of Unilateral Action." In The Oxford Handbook of the American Presidency, eds. George C. Edwards III and William G. Howell. Oxford, UK: Oxford University Press, 427-54.

Mayer, Kenneth R., and Kevin Price. 2002. "Unilateral Presidential Powers: Significant Executive Orders, 1949-99-" Presidential Studies Quarterly 32 (2): 367-85.

Mayhew, David R. 1991- Divided We Govern. New Haven, CT: Yale University Press.

Milkis, Sidney M., and Jesse H. Rhodes. 2007. "George W. Bush, the Republican Party, and the 'New' American Party System." Perspectives on Politics 5 (3): 478-503.

Moe, Terry M., and William G. Howell. 1999- "Unilateral Action and Presidential Power: A Theory." Presidential Studies Quarterly 29 (4): 850-73.

Neustadt, Richard. [1960] 1991. Presidential Power and the Modern Presidents. New York: The Free Press.

Obama, Barack H. 2009. "Mexico City Policy and Assistance for Voluntary Population Planning." Federal Register 74: 4903-04.

"Obama Swiftly Reverses Bush Orders." 2009. Nature 457 (7229): 523.

Ostrander, Ian, and Joel Sievert. 2013. "The Logic of Presidential Signing Statements." Political Research Quarterly 66 (1): 141-53.

Reagan, Ronald W. 1981a. "Federal Civilian Employee Hiring Freeze." Federal Register 46: 9907.

--. 1981b. "Postponement of Pending Regulations." Federal Register 46: 11227.

Rottinghaus, Brandon, and Jason Maier. 2007. "The Power of Decree: Recent Presidential Use of Executive Proclamations." Political Research Quarterly 60 (2): 338-43.

Rudalevige, Andrew C. 2012. "Executive Orders & Presidential Unilateralism." Presidential Studies Quarterly 42 (1): 138-60.

--. 2014. "The Unilateral President? An Addendum." The Monkey Cage, January 29- http:// (accessed January 31, 2014).

Savage, Charlie. 2006. "Bush Challenges Hundreds of Laws." Boston Globe, April 30. http:// bush_challenges_hundreds_of_laws/ (accessed January 21, 2014).

--. 2009. "Obama's Embrace of a Bush Tactic Riles Congress." New York Times, August 8. (accessed January 21, 2014).

--. 2010. "Obama Takes New Route to Opposing Parts of Laws." New York Times, January 8. /09/us/politics/ 09signing.html?scp=4&sq=charlie%20savage&st=cse&_r=0 (accessed January 21, 2014).

Tapper, Jake, Sunlen Miller, and Huma Khan. 2009. "Obama Overturns 'Mexico City Policy' Implemented by Reagan." ABC News, January 23. story?id=6716958 (accessed July 15, 2014).

Warber, Adam L. 2006. Executive Orders and the Modern Presidency: Legislating from the Oval Office. Boulder, CO: Lynne Rienner Publishers.

Wolf, Richard. 2011. "Obama Uses Executive Powers to Get Past Congress." USA Today, October 27. 50942170/1 (accessed July 18, 2014).

(1.) Throughout the article, I use the phrase "unilateral action" to indicate presidential behavior undertaken to alter the status quo of some policy. Of course, there are important questions surrounding this term. For instance, whether the outputs researchers observe are, indeed, the revealed preferences of presidents--and whether the actions taken actually succeed in altering the status quo. Rudalevige (2012) considers the former question in the context of executive orders, finding that archival evidence suggests the formulation of those actions is not "unitary" in any sense. To be clear, I do not argue that in every case, memoranda and orders satisfy a stricter definition of unilateral action. I proceed with the more modest assumption that they are, at the very least, indications of an underlying propensity toward "acting alone"--with eventual policy outcomes remaining uncertain.

(2.) This finding is largely consistent with Cooper (2002) and Dodds (2013), who argue persuasively that the substantive significance and frequency of these informal, alternative means of presidential action has increased.

(3.) To illustrate this point, a Google search of recent news articles (on January 31, 2014) using the keywords "executive orders" yields roughly 110,000 results, whereas "presidential memoranda" yields 3,300. Naturally, this disparity could be a function of differences in the salience of the issues addressed by these tools or the fact that memoranda are sometimes misreported as executive orders. Many articles, however, address the use of executive orders in general--a signal of elite and mass public interest in the topic.

(4.) Aberbach and Rockman make a similar point, writing, "presidents will do what they can to push their preferences and find loopholes to circumvent a recalcitrant Congress''(1999, 757).

(5.) I add this modifier because it is clear some presidential directives involve negotiation with relevant stakeholders and administrators.

(6.) Krause and Melusky (2012) find that unilateral actions give executives the capacity to pursue "short-term" electoral gains. Memorandum issuance ought to be consistent with these pursuits.

(7.) These ordered reviews of federal programs and eventually led to the executive reorganization of 1978.

(8.) Reagan issued the two memoranda at the outset of his first term--"Federal civilian employee hiring freeze" and "Postponement of pending regulations" (Reagan 1981a, 1981b). Though retrospective accounts question the real impact of the hiring freeze, the administration would later go on to claim that the freeze was eliminating five hundred employees per day from the federal payroll (Ayres 1981).

(9.) The memoranda says, in part, "Because Puerto Rico's degree of constitutional self- government, population, and size set it apart from other areas also subject to Federal jurisdiction under Article IV, section 3, clause 2 of the Constitution, 1 hereby direct all Federal departments, agencies, and officials, to the extent consistent with the Constitution and the laws of the United States, henceforward to treat Puerto Rico administratively as if it were a State" (Bush 1992; emphasis added). This case is noteworthy, because this area of policy typically had been made via executive order. Franklin Roosevelt issued Executive Order 7057 in 1935, which carried out Puerto Rican relief programs legislated by Congress. Richard Nixon issued another (Executive Order 11755) pertaining to Puerto Rico's prisons in 1973.

(10.) President Obama issued a total of 162 memoranda and 171 executive orders from 2009 to 2013.

(11.) "Transparency and Open Government," January 26, 2009 (Federal Register vol. 74, p. 4685); "Pay Freeze," January 26, 2009 (Federal Register vol. 74, p. 4683); "Expediting Review of Pipeline Projects From Cushing, Oklahoma, to Port Arthur, Texas, and Other Domestic Pipeline Infrastructure Projects," March 3, 2012 (Federal Register vol. 77, p. 18891); "Flexible Implementation of the Mercury and Air Toxics Standards Rule," December 27, 2011 (Federal Register vol. 76, p. 80727); "Freeze on Discretionary Awards, Bonuses, and Similar Payments for Federal Political Appointees," August 6, 2010 (Federal Register vol. 75, p. 47433); "Establishing a Task Force on Childhood Obesity," February 18, 2010; "Task Force on Space Industry Workforce and Economic Development," May 5, 2010 (Federal Register vol. 75, p. 24781); "Establishing an Interagency Task Force on Federal Contracting Opportunities for Small Businesses," April 29, 2010 (Federal Register vol. 75, p. 22499); "White House Task Force on Middle-Class Working Families," February 3, 2009 (Federal Register vol. 74, p. 5979).

(12.) The digitized Federal Register contains a comprehensive list of published memoranda from 2000, whereas the Code of Federal Regulations was used for years prior. See (accessed July 18, 2014).

(13.) The Code of Federal Regulations is available electronically from 1996. Earlier listings were provided by request from the National Archives and Records Administration. See collectionCfr.action?collectionCode=CFR (accessed July 18, 2014).

(14.) National Archives Identifier: 3438862; HMS number: A1 36.

(15.) The last "letter" in my data was issued by Carter on April 22, 1978; Federal Register vol. 43, p. 17789.

(16.) This equivalence is apparent when one examines each action by function, rather than by title. One of the most readily identifiable functions of memoranda have been delegations of authority. This particular function appears in both "letters" (April 27, 1959 [Federal Register vol. 12, p. 3691]; June 2, 1970 [Federal Register vol. 35, p. 8631]) and memoranda (October 3, 1980 [Federal Register vol. 45, p. 67629]; February 3, 1993 [Federal Register vol. 58, p. 8203]).

(17.) A determination communicates to Congress presidential findings, which are most often statutory requirements of trade agreements. During the Clinton presidency, determinations were numbered by year, and then, sequentially within each year. For example, Presidential Determination No. 93-22 was the twenty-second determination of 1993.

(18.) These records have only been digitized as far back as 1993, making a substantial proportion of the memoranda time consuming to collect.

(19.) Many of the actions reflect partisan goals, which is largely consistent Milkis and Rhodes' (2007) analysis of the "new" executive-centered party system.

(20.) As Mayer (1999) notes, this requirement began in earnest after the Federal Register Act (44U.S.C. [section] 1505).

(21.) Cooper (2002) tabulates the number of memoranda from 1993 to 2000 using the Weekly Compilation. He finds 536 memoranda issued during that period. By contrast, my data contain 108 in that time period. It is worth noting, however, that many of the cases Cooper examines were published in the Federal Register, and thus, are contained in my data.

(22.) Several news outlets [ABC, Nature) falsely reported Obama's "Mexico City" memoranda as an executive order; see "Obama Swiftly Reverses Bush Orders" (2009); Tapper, Miller, and Khan (2009).

(23.) Warber (2006) analyzed the content of executive orders through a process designed to determine significance. Such a procedure might be employed in the case of presidential memoranda.

(24.) (accessed July 15, 2014).

(25.) This marginal effect is identical for both variable specifications.

(26.) As one reviewer pointed out, the rise in regulation and a corresponding need for presidents to control the process may also help explain the increase. This explanation applies to a subset of the memoranda considered. Given that memoranda perform a variety of functions, a complete account of their use has to consider the diversity of incentives associated with each of those functions.


University of Virginia

Kenneth S. Lowande is a Ph.D. student in the Department of Politics at the University of Virginia. His research interests include bureaucracy, the presidency, and legislative-executive relations.

Marginal Effects (at Means) on Memorandum Use

Variable                           Memoranda       Memoranda (alt.)

Executive [Orders.sub.AR(2)]    0.107 (0.051) **    0.090 (0.04) **
Executive [Orders (alt.)
Divided Government              -4.95 (1.71) ***    -5.09 (1.45) ***
Beginning of Term               -1.61 (1.26)        -2.12 (0.98)
Approval Rating                -0.056 (0.06)       -0.026 (0.051)
Party Change                    -2.60 (2.15)        -1.62 (1.65)

Note: full model reported in Table 2.

Standard errors in parentheses; *** p < .01, ** p < .05, * p < .1
Presidential dummy variables and lagged-memoranda omitted for

Patterns of Memoranda Use, 1946-2013; Negative Binomial
(Linear Model in Equation 1). Executive Orders Modeled
as AR(2) Process to Remove Time-Dependent Variation

Variable                                     Memoranda        (alt.)

Executive [Orders.sub.AR(2)]                   0.0103 **     (0.005)
Executive [Orders (alt.).sub.AR(2)]                           0.0144 **
Divided Government                            -0.476 ***     -0.812 ***
                                              (0.16)         (0.22)
Beginning of Term                             -0.155         -0.338 **
                                              (0.12)         (0.15)
Approval Rating                               -0.005         -0.004
                                              (0.006)        (0.008)
Party Change                                  -0.251         -0.258
                                              (0.21)         (0.26)
President                    Obama (Base)
                             W. Bush          -0.813 ***     -1 757
                                              (0.22)         (0.33)
                             Clinton          -0.713 ***     -1.144 ***
                                              (0.21)         (0.27)
                             H. W. Bush       -0.798 ***     -1.178 ***
                                              (0.28)         (0.36)
                             Reagan           -0.779 ***     -0.859 ***
                                              (0.25)         (0.28)
                             Carter           -1.145 ***     -1.070 ***
                                              (0.27)         (0.29)
                             Ford             -1.012 **      -0.830 **
                                              (0.39)         (0.41)
                             Nixon            -2.213 ***     -2.141 ***
                                              (0.39)         (0.42)
                             Johnson          -2.828 ***     _3 324
                                              (0.43)         (0.52)
                             Kennedy          -1.567 ***     -1.562 ***
                                              (0.36)         (0.42)
                             Eisenhower       -1.978 ***     -1.750 ***
                                              (0.34)         (0.41)
                             Truman           -2.303 ***     -2.278 ***
                                              (0.39)         (0.41)
[Memoranda.sub.t-1]                            0.008
[Memoranda (alt.).sub.t-1]                                   -0.003
Constant                                       3.976 ***      4.027 ***
                                              (0.50)         (0.53)
Log-likelihood                              -172.10        -154.85
[chi square] (16)                            106.01          76.85
Pseudo-[R.sup.2]                               0.236          0.199
N                                             67             67

Standard errors in parentheses; an "alt." variable indicates
designations, delegations, and certifications have been
removed from the yearly count. *** p < .01, ** p < .05.
COPYRIGHT 2014 Wiley Periodicals, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2014 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:The Contemporary Presidency
Author:Lowande, Kenneth S.
Publication:Presidential Studies Quarterly
Article Type:Report
Geographic Code:1USA
Date:Dec 1, 2014
Previous Article:Appointee-careerist relations in the presidential transition of 2008-2009.
Next Article:Public perceptions regarding the authenticity of the 2012 presidential candidates.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters