“[I]f you want to get into a debate with a four-star Marine general, I think that that’s something highly inappropriate.”
Sarah Sanders, White House Press Secretary1
The United States has experienced multiple episodes of civil-military tension that have bordered on eroding norms of civilian control over the last 30 years.2 The Trump administration amplified the politicization of the military to the degree that it became a distinct feature of civil-military relations during former President Donald Trump’s presidency.3 Specifically, the increased reliance on the military in policymaking became a salient feature of U.S. civil-military relations during the Trump administration and is likely to have lasting consequences. Trump’s initial set of appointees included Gen. (Ret.) James Mattis as secretary of defense, Gen. (Ret.) John Kelly as secretary of homeland security, and Lt. Gen. (Ret.) Michael Flynn as national security adviser. Even before Trump’s inauguration, scholars of civil-military relations voiced concerns about the growing influence of former military members in his administration.4 It did not seem, however, that the new president shared this concern. Regardless of the retired or active-duty status of his team members, Trump referred to them as “my generals,” underscoring their connection to the military profession and institution.
Trump’s intention to invoke the Insurrection Act of 1807 and use the military in response to the Black Lives Matter protests in June 2020, as well as the subsequent statements of Mattis, Adm. (Ret.) Michael Mullen, and Gen. (Ret.) Martin Dempsey condemning this move, once again brought to the fore the increased role that the armed forces play in U.S. politics today.5 Most importantly, this example showed that high-ranking members of the military profession deemed it appropriate, and even necessary, to use their political power to impose constraints on presidential policy decisions. These developments warrant a thorough examination of how the extensive reliance on military elites in policymaking affects civilian control and democratic governance.
The erosion of civil-military norms is not a unique aberration that will fix itself with the conclusion of Trump’s presidential term. While Trump is not the first U.S. president to rely on senior military officers in policymaking, his presidency was characterized by what Loren DeJonge Schulman calls a “hurricane of civil-military norm upending.”6 President Joe Biden’s team will have to deal with the damage left by that storm. Despite the fact that the United States now has a new administration, the U.S. military carries the legacy of erosion of civil-military norms. Moreover, the societal perception of what is and what is not an appropriate degree of military influence on policy has also shifted in the last four years. Thus, the Biden administration will have to remedy both the erosion of civil-military norms and the resulting shift of bureaucratic patterns. In order to do this, it is critical to understand the drivers behind the executive’s delegation of policymaking prerogatives to members of the military profession within the Trump administration and beyond it. In addition, identifying what is behind the increased reliance on the military in policymaking is crucial for recognizing the risks for civilian control and preventing the future weakening of democratic norms.
For a long time, scholarship on civil-military relations has focused on the erosion of civilian control originating from the military, including coups, defections, and insubordination.7 Unsurprisingly, the most studied form of erosion of civilian control is a coup.8 However, the military is not the only actor with the agency to shift the balance of power and weaken civilian control.9 The government’s own actions can decrease the power of civilians in policymaking. Presumably, the civilian government is interested in precisely the opposite — limiting the military’s political power and exercising robust civilian control.10 However, recent and historical events point to the contrary and indicate an increased reliance on the military in policymaking. This observation motivates this paper’s central question: Why would democratically elected officials voluntarily delegate part of the authority entrusted in them by the voters to members of the military?
To answer this question, I first introduce a new concept — the erosion of civilian control of the military by the executive’’s deference to members of the military profession in policymaking (henceforth: erosion by deference). Conceptualizing this relevant yet understudied phenomenon will help to capture the decrease of relative civilian power in policymaking occurring through the intentional delegation of authority to the military by elected officials. Then, relying on the literature on civil-military relations, regime legitimacy, and democratic governance, I propose three drivers of deference — boosting popular approval, avoiding responsibility, and cajoling the military. I also specify the conditions that enable or spark certain mechanisms of erosion, including high and growing popular trust in the military, low and decreasing popular approval of the executive, the executive’s departure from previously stated policy goals, high civil-military tensions, and a degree of convergence between civilian and military preferences.
Using detailed case studies from the Trump, Clinton, and Bush (43) administrations, I demonstrate the analytical utility of the new concept of erosion by deference and its drivers. The Trump administration case provides a data-reach research context in which most of the conditions making erosion by deference likely are present. Therefore, studying this case allows for the observance of all three abovementioned drivers of deference while keeping historical, cultural, and personalistic variables constant. The addition of the Clinton and Bush administrations shows that, while the degree and scope of erosion by deference under Trump was unprecedented, this phenomenon is produced by factors outside just the executive’s personality.
The article then analyzes the qualitative and quantitative evidence for the erosion of civilian control by deference under the Trump administration, showing how delegating policymaking tasks to the military weakened civilian control. The concluding discussion suggests that erosion of civilian control by deference is relevant outside the United States and is especially alarming for states with high societal trust in the military, populist leadership, and an overall tendency for democratic backsliding.
Why Is Deference a Problem?
Before answering the central question of this study, it is important to discuss why civilian deference to the military in policymaking is problematic, especially in democratic regimes. To begin, civilian control of the military is a necessary attribute of democratic governance and a crucial step in democratization.11 The actors that control the military in democracies — be they the executive, the legislature, or both — acquire legitimacy through public participation in elections.12 In addition, besides preventing coups and assuring subordination, democratic civilian control requires civilian dominance in policymaking,13 including relying on civilian expertise in national security and foreign policy.14 Thus, while civilian dominance in policymaking is optional for autocracies, it is an essential requirement in democracies.
The concept of erosion by deference challenges the idea that delegating policymaking responsibilities to members of the military profession does not constitute a threat to civilian control.15 The argument that delegation does not constitute a problem for civilian control rests on the assumption that civilians can decide when, how, and how much power to give to the military and that they can take it back whenever they see fit. But this argument does not consider the third actor in this power relationship: the citizen. In democratic states, political power is conditioned on popular support of the government and societal trust in institutions. Since power does not wholly reside in elected officials and depends on popular approval, civilian officials may not be free to take back fully the power that they have delegated. In particular, they cannot reverse popular feelings toward members of the military that the military’s increased participation in policymaking will engender. For instance, popular acceptance of the military as an authority in policymaking and tolerance of the military’s participation in partisan politics may not be easily reversed by appointing a civilian in a policymaking role that was previously held by a member of the military. Furthermore, military officers charged with fulfilling political duties may acquire a sense of entitlement and superiority and lose respect for civilian authorities who delegated their responsibilities to the uniformed services.16
The concept of erosion by deference challenges the idea that delegating policymaking responsibilities to members of the military profession does not constitute a threat to civilian control.
More specifically, if the policies that the military introduces are successful, delegation can increase the overall political influence of the military — an institution that should not be involved in the political power competition in democracies at all. Increased political credibility of the military may enable military officers to impose constraints on civilian executives’ decision-making by publicly claiming approval or disapproval of certain policies. Moreover, delegating the responsibility for policymaking to the current and former members of the military undermines the perception of the armed forces as a non-partisan institution.17 On the other hand, if the policies that the members of the military profession introduce are unsuccessful, it can cast a shadow of popular mistrust on an institution whose members, unlike politicians whose policies were unsuccessful, cannot be reelected, undermining faith in a critical public institution. In turn, if citizens consider the armed forces to be a partisan player, popular trust in the military becomes a function of party affiliation. Such an arrangement could negatively affect recruitment to the armed forces, discouraging the supporters of certain political parties from joining. This connection between service and partisanship would make the military less representative of the population, which is critically important in democracies where popular trust in the military protecting the state and all members of its society is crucial. The military’s partisanship could be especially detrimental in states with an all-volunteer force like the United States.
Of course, some argue that it is impossible and, in fact, undesirable to separate the military from policymaking because strategy and policy are inherently interconnected.18 Nevertheless, the delegation of policymaking responsibilities to members of the military profession is different from consulting with military experts on matters of policy. In what Eliot Cohen calls an “unequal dialogue” between civilian executives and the military, “both groups must expect a running conversation in which, although civilian opinion will not dictate, it must dominate.”19 This is why arguing that civilians lack expertise in certain areas and need the military’s input on policy matters only partially explains the delegation of policymaking prerogatives to the military.20 Specifically, while soliciting military advice to inform the evaluation of threats and opportunities or ways and means is warranted, charging officers with the responsibility to decide on these matters undermines civilian dominance in policymaking.
In addition, empirical studies on the effect of having a military background on decision-making support the existence of what Samuel Huntington calls the “military mind.”21 In particular, political elites with a military background have divergent policy preferences compared to their civilian counterparts, especially when it comes to the use of force.22 Some members of the military profession may have preferences that would actually strengthen civilian oversight.23 However, the extensive participation of the military in government might increase the reliance on the use of force in foreign policy.24 Moreover, even in democratic countries, the higher ranks of the military (which includes those who are most likely to get involved in politics) are not diverse or inclusive. For example, in the United States, women, people of color, and ethnic and religious minorities tend to be underrepresented in the higher echelons of the military profession.25 Therefore, heavy reliance on the military in policymaking can amplify biases in decision-making and limit the input of underrepresented communities in politics.
It is important to note that erosion by deference is different from other forms of weakening civilian control (i.e., erosion by insubordination or competition) because it originates from the actions of the civilian executive, which are the main focus of this paper.26 When elected executives delegate policymaking responsibilities to members of the military profession, they do so because they are pursuing certain political goals that someone with a military background might help to accomplish. For example, appointing a general in a top policymaking position could increase popular confidence in the government’s policies on a particular issue. At the same time, it could limit civilian input in the policy process, leading to erosion of civilian control.27 It is of less importance whether retired generals are technically civilians so long as the civilian executive is taking advantage of their military credentials. For this reason, the civilian executive’s motivations for deference constitute the core concern of this research.
In sum, the delegation of policymaking tasks to members of the military profession can decrease civilian input in politics, elevate the political weight of military elites, undermine trust between civilian executives and the military,28 question the non-partisan nature of the armed forces, undermine popular confidence in them, and introduce biases to the decision-making process. These developments can have a devastating effect not only on civilian control but also on other democratic institutions and processes. Thus, understanding the drivers of the erosion of civilian control by deference is critical for acquiring a more nuanced perspective of the broader processes of democratic backsliding.29
Defining Civilian Deference to the Military
Although academic literature and commentary discuss the phenomenon of civilian deference to the military and its erosive potential for civilian control, the concept itself still lacks a definition.30 The definition I propose fills the conceptual gap and prepares the ground for further research. I define erosion of civilian control by deference to the military as a shift in the civil-military power balance, in which civilian authorities delegate policy tasks typically fulfilled by civilian actors to members of the military profession.
Erosion of civilian control refers to the weakening of civilian power relative to the military when compared to existing norms or a previous period of observation. For instance, if having a retired chief of the general staff as a minister of defense in Israel is the norm, appointing another retired general to this position would not constitute the erosion of civilian control by deference, but would rather signify the inherent weakness of civilian control. Alternatively, in the United States, which has a long-held tradition of civilian dominance in the Department of Defense, appointing a secretary of defense who has a prominent military background and brings in a predominantly uniformed team would constitute an aberration from the status quo.31
The colloquial meaning of deference implies trust, respect, and awe. This trust, respect, or awe on the side of civilian leaders can be genuine or performative. In other words, it is not important whether civilian executives really trust the military, say they trust the military, or behave as if they trust the military with regard to policy. Most important is the delegation of typically civilian prerogatives (as defined by law or recent custom) to current or former military members. This delegation might include the official appointment of a member of the military profession to a policymaking position that is usually held by a civilian, the assignment of policy tasks that are usually performed by civilian politicians to a military member, or the withdrawal of civilian executives from the policymaking process. Thus, the delegation of power can happen either by omission or commission.
[E]rosion of civilian control of the military is not the initial aim of civilian executives, but rather is a byproduct of their efforts to attain other political goals.
The above definition uses the broad term “members of the military profession,” which includes active-duty military officers and recently retired individuals who maintain close connections to the military institution, are publicly perceived as representatives of the armed forces (e.g., referred to as “general” or “colonel”), appear in uniform in public, or self-identify in terms of their military service.32 Civilian executives referring to retired military members in their government by their rank contributes to blurring the boundaries between civilian appointees and the military. A broad conceptualization of the military is necessary to account for the fact that the actions of recently retired members of the military profession still reflect on the popular image of the armed forces and impact the military’s power to affect policy.33 In addition, empirical research shows that service in the military has a consistent long-term effect on an individual’s policy preferences that does not disappear with retirement from the armed forces.34
The policy tasks that civilian officials delegate to the military can include both policy formulation and implementation. Formulation involves assigning the primary responsibility for identifying policy objectives, evaluating limitations and opportunities, deciding on the ways and means, or selecting the preferred course of action. Implementation includes tasks usually fulfilled by civilians: negotiating with adversaries and partners, performing diplomatic and representative duties, etc.
Drivers of Deference
Turning to this paper’s central puzzle, why would civilian authorities elected by a popular vote voluntarily delegate their policymaking power to those in uniform?
The first theoretical assumption that will help to identify the drivers of deference is that erosion of civilian control of the military is not the initial aim of civilian executives, but rather is a byproduct of their efforts to attain other political goals. Understanding this leads to the second theoretical assumption: Delegating power to the military has to be more beneficial for the government than keeping it, in order for erosion by deference to take place. To explain why the government would find it beneficial to voluntarily delegate political power to the military, one must take into account the power relationships between the civilian population, the government, and the military.35 The government acquires the legitimacy to rule and to control the military when the citizenry delegates this authority to government officials through democratic elections.36 Citizens also endow the armed forces with a certain level of confidence that rests on the societal belief that the military will protect the state and its citizens when necessary. The combination of popular support for the government, healthy civil-military relations, and citizens’ confidence in the armed forces allow the executive to exercise power with minimum friction.
I argue that the government will defer to the military in policymaking to secure higher support either from the military or the citizenry in order to maintain or increase its power.37 Popular confidence in the military (if present) provides it with political capital that the government could exploit to win more popular support for a given policy. Alternatively, the government could use the military in policymaking to avoid taking responsibility for costly political moves and to prevent or minimize the loss of popular approval. Finally, if civilian executives want to alleviate or prevent civil-military tensions, they can invite some officers into the policymaking process to bring harmony to the relationship with the military. These three possible scenarios translate into three drivers of the government deferring to the military: boosting approval, avoiding responsibility, and cajoling the military with power.
A potential alternative explanation available in the literature on civil-military relations suggests that elected politicians delegate policymaking prerogatives to the military when civilians lack the necessary expertise in defense matters.38 This explanation, though plausible in some cases (e.g., Israel), does not explain the delegation of policymaking powers to the military in the United States, where extensive civilian expertise in defense is established and available.39 Moreover, the expertise-based explanation falls short of illuminating the deep-rooted motives and conditions that prevented the development of civilian defense expertise in cases where this explanation does apply.40 Thus, the discussion below of the drivers of deference goes beyond questions of expertise and explores the motivations for delegation that are entrenched in the triangular structure of the civil-military power balance.
Boosting Approval
Civilian officials may choose to invite the military into the policymaking process to increase popular support for the government or its particular policies. For this driver of deference to work, the military has to enjoy high levels of confidence from the public. In many democratic countries such as France, the United Kingdom, Denmark, and Germany the military has higher popular trust than other institutions.41 Existing research also shows that, in the United States, public confidence in the military has been consistently high in recent decades as opposed to trust in civilian institutions.42 This makes the military not only a capable tool of statecraft but also a potentially valuable political asset.43 Indeed, empirical findings indicate that the attitudes of military elites about the use of force have strong effects on public support for particular missions.44 Moreover, a recent study shows that the American public is willing to defer to the military’s opinion not only on matters of security and defense but also foreign policy.45 Therefore, sharing the responsibility for policymaking with members of the military profession can increase public support for government policies.
An additional condition for civilian deference in order to boost approval is low popular support for the government or its policies. The government may use popular confidence in the military to increase its own public approval. If attempting to boost approval drives civilian deference to the military, we would likely observe civilian leadership emphasizing the military’s support for its policies in public statements, civilian leaders and military members appearing together in press conferences, or civilian officials making a military officer point person for a controversial policy.
For instance, Trump’s appointees early in his presidency included a remarkably high number of members of the military profession in positions with extensive policy influence — Mattis as secretary of defense, Kelly as secretary of homeland security, and Flynn as national security adviser (later replaced by active-duty Lt. Gen. H. R. McMaster). Of course, having a retired general as a secretary of defense is not unprecedented. However, the last time such a high-ranking officer served as secretary of defense was Gen. George Marshal in 1951 — 66 years before Mattis’ appointment. Moreover, out of the 11 defense secretaries in the post-Cold War era, only two were career military officers (Mattis and Mark Esper — both under Trump’s presidency), and only one of them (Mattis) held the rank of general. Appointing Mattis required special congressional approval because he retired from the military only three years before the appointment, not seven as required by law. Thus, his appointment constituted an aberration from existing norms, consistent with the definition of erosion by deference. It is important to note that erosion of civilian control does not come from appointing generals to these positions but from the civilian executive taking advantage of those individuals’ military background. For instance, Trump persistently called the secretaries of state and homeland security “my generals,” which alarmed scholars of civil-military relations and the military.46
Almost immediately after entering office, Trump used the credibility of “his generals” to promote controversial policies discussed during his campaign. Executive Order 13767on “Border Security and Immigration Enforcement Improvements” included building a physical wall on America’s southern border.47 Executive Order 13769 on “Protecting the Nation from Foreign Terrorist Entry Into the United States” introduced the policy of “suspension of issuance of visas and other immigration benefits to nationals of countries of particular concern,” colloquially known as the “travel ban.”48 While both orders fell outside the Department of Defense’s purview, Trump signed them in the Pentagon’s Hall of Heroes, with Mattis by his side.49 This move immediately sparked speculation that the president was using the credibility and trust in the U.S. military to increase support for his divisive policies.50
It is important to note that erosion of civilian control does not come from appointing generals to these positions but from the civilian executive taking advantage of those individuals’ military background.
Indeed, Mattis was not merely a member of the military profession, but one of the most revered representatives of the armed services, respected both by Democrats and Republicans.51 At the same time, Trump’s net approval was considerably lower than of any of his predecessors, making him the only American president with net approval so close to zero at the beginning of his term since the beginning of the Cold War.52 The combination of low popular support, strong preferences for ambitious and divisive policies, and projecting the image of the Department of Defense’s support for his policies is consistent with the approval booster explanation.
The Trump administration is not unique in relying on highly respected military officers to compensate for a lack of policy approval. When President George W. Bush’s Iraqi surge policy was lacking support among the public and in Congress, his administration instituted a new position of assistant to the president and deputy national security adviser for Iraq and Afghanistan. This new post, colloquially referred to as “the war czar,” was offered only to prominent military officers, three of whom — all highly respected four-star generals — refused the position.53 Lt. Gen. Douglas Lute, who accepted the offer, described his responsibilities as “developing policy adaptations to meet changing needs on the ground.”54 In fact, he became responsible for coordinating the interagency policy efforts amid harsh congressional and public pressure to revise the U.S. strategy in Iraq.
Appointing a three-star general to this unusual position created several civil-military complications apparent from the very beginning, many of which were voiced in Lute’s congressional confirmation hearing. First, Sen. John Warner mentioned that it might be problematic that Lute would have to manage the efforts of high civilian officials from various U.S. governmental bodies. Second, Sen. James Webb pointed out the civil-military relations problem that inevitably occurs “when you have a uniformed military individual making political judgments and giving political advice to a political administration.”55 Third, the appointment of a highly respected active-duty army general to a newly created position that had significant overlap with already existing posts called into question the competence of senior civilian politicians. In particular, Sen. Jack Reed boldly acknowledged to Lute that “all you’re being asked to do was what Stephen Hadley and Dr. Condoleezza Rice were supposed to be doing for the last several years.”56 While raising these problems with the appointment, the senators also admitted their highest respect for Lute and his service and confirmed him as a new “war czar.”
What motivated offering this unusual position exclusively to members of the military profession? The circumstances at the time point to the boosting approval explanation. Bush had a strong policy preference for initiating the surge, despite knowing from the very beginning that popular and congressional support for the new strategy in Iraq was low.57 This combination of a strong policy preference and low support is consistent with the approval booster explanation. Public opinion polling indicates that in April 2007, when the Bush administration was creating the new “war czar” position, 41 percent of respondents claimed that the surge was not making any difference, 29 percent stated that it was making the situation even worse, while only 26 percent thought that the surge was improving the situation.58 Congress reflected the popular sentiment. Not only Democratic but also some prominent Republican members of Congress criticized Bush’s policy in Iraq and demanded a change of course.59 To buy some time to allow the surge to bear fruit, the president needed to increase popular and congressional confidence in the policy effort.
The non-partisan image of the military and the high respect that military officers command on both sides of the aisle made Lute an ideal intermediary between the president and Congress. Indeed, in July 2007, when congressional tensions intensified, the Bush administration sent Lute to Capitol Hill to convince the senators to wait for Gen. David Petraeus and Amb. Ryan Crocker’s September reports on the progress of the surge before demanding a change of policy.60 Although such lobbying by members of the presidential team does not break any conventions, an active-duty general advocating for a partisan policy preference on the conduct of war does. This use of a respected military officer to galvanize support for an unpopular policy is consistent with the boosting approval explanation for deference.
Could the other drivers of deference also be at play in this example? Because Bush did not try to distance himself from the surge policy, this example does not meet the necessary criteria for the explanation that he was seeking to avoid responsibility. For instance, in the 2007 State of the Union address, Bush takes full responsibility for the political decision to increase the number of troops in Iraq, saying, “I chose this course of action because it provides the best chance for success.”61 Bush also engaged in an open fight over the Iraq policy with Congress in the summer of 2007.62 The president’s surprise visit to Iraq in September 2007 made the news and clearly showed that he was not trying to distance himself from the situation in Iraq.63 In addition, the initial refusal of the three four-star generals to accept this appointment casts doubt on the cajoling the military explanation, which requires that the military want to exercise more power over the policy in question. Taking this into account, it is most likely that offering the responsibility for an interagency policy effort exclusively to senior members of the military profession was supposed to mitigate the lack of support for Bush’s Iraq policy.
In the above examples, the senior executives who delegated the policymaking responsibilities to the members of the military profession were motivated by the lack of public approval of their preferred policies. It is important to note that it is unlikely that by delegating the authority to members of the military profession, Trump and Bush were trying to weaken civilian control. However, the predictable downside of using the military to increase approval is that it limits civilian input in policymaking and increases the military’s political weight.
Avoiding Responsibility
The second driver of civilian deference to the military is avoiding responsibility. In this case, delegating policymaking responsibilities to the military constitutes a benefit in itself because it allows the government to avoid the political costs associated with the policy. While civilian authorities do have the right to be wrong,64 the price for exercising this right can be losing public approval and reelection. Therefore, when forced to make difficult political decisions, the government may abdicate its policymaking responsibilities and delegate them to military officers.65 James Golby and Mara Karlin describe this phenomenon as “using military leaders to shield elected officials from political criticism.”66 They also discuss how this practice increases the relative bargaining power of the military in policymaking.
When avoiding responsibility drives erosion by deference, the government will demonstrate reluctance to proceed with the policy in question. Unlike with boosting approval, where elected officials present a joint front with members of the military profession to gain more support, the key objective of deference driven by avoiding responsibility is not to lose support. Therefore, the executive has to distance himself or herself from the policy. This driver of deference is most likely to be set in motion when the policy bears potential costs to the elected executive. Such a policy might involve undermining campaign promises or reversing previously promoted policies, a high likelihood of casualties, a rapid increase in spending, or cooperating with a political opponent. In practice, the executive is likely to withdraw from the policymaking process and even blame the military for adopting the policy.
For example, on July 13, 2017, Trump delegated the authority to determine troop levels in Afghanistan to then-Secretary of Defense Mattis. This decision expanded the Defense Department’s responsibility to an unprecedented level.67 While the military can make recommendations about the required number of troops, only elected civilian officials can take responsibility for sending more U.S. citizens to war.68 In fact, the president of the United States has always made this decision. Leaving this choice to the full discretion of the military-dominated Department of Defense decreased civilian input in this crucial policy matter. Therefore, this move matches the definition of erosion by deference.
What explains Trump’s decision to delegate power to Mattis and the Defense Department? The evidence indicates that avoiding responsibility is a plausible explanation. First, increasing troop deployments contradicted Trump’s policy preferences and campaign promises. Specifically, during his foreign policy speech in April 2016, candidate Trump received a loud round of applause for saying, “I will never send our finest into battle unless necessary, and I mean absolutely necessary, and will only do so if we have a plan for victory with a capital V.”69 He also continually criticized previous administrations for spending money on the longest war in American history.70 As president, Trump continued to be reluctant to increase the number of American troops in Afghanistan.71 Meanwhile, both the commander of U.S. troops in Afghanistan and the secretary of defense admitted that the United States was not winning the war, casting doubt on Trump’s chances to deliver a victory with a capital V to his supporters.72
When avoiding responsibility drives erosion by deference, the government will demonstrate reluctance to proceed with the policy in question.
The president made several moves to distance himself from this potentially costly policy. First, he delegated the difficult decision to the Department of Defense. Then he publicly stated his reluctance to follow through with the proposed plan. Specifically, when Trump announced the revised strategy in Afghanistan in August 2017, he informed the audience, “I share the American people’s frustration. I also share their frustration over a foreign policy that has [spent] too much time trying to rebuild countries in our own image instead of pursuing our security interests.”73 Trump also insisted that the decision was against his instincts, which he was convinced not to follow this time. Such behavior suggests that avoiding responsibility was the primary driver for erosion by deference in this case. In addition, distancing himself from the military and the decisions it made excludes the boosting approval explanation, which presumes a visible joint civil-military policy effort.
Delegating policymaking prerogatives to members of the military profession in the face of costly policies or challenging crises is not unique to the Trump administration. This was what happened when President Bill Clinton appointed Gen. Barry McCaffrey as director of the Office of National Drug Control Policy in 1996. When Clinton decided to appoint a new “drug czar,” his domestic-policy aide Rahm Emanuel suggested four options: a crime-buster, a school principal, a prosecutor, or a soldier. Clinton interviewed only one person — McCaffrey, who was a four-star general.74 Before McCaffrey, only civilians had occupied the position of drug czar.75
Why did Clinton decide to appoint a four-star general to this position? The evidence suggests that avoiding responsibility is the most likely explanation. First, falling popular approval over the administration’s sluggish anti-drug policy was threatening Clinton’s reelection. Indeed, the public, media, and Congress were critical of the administration’s lack of a consistent strategy, visible efforts, and tangible positive changes in fighting illegal drugs. Clinton’s previous drug czar, former New York City police commissioner Lee Brown, was ridiculed in the press and within the administration.76 In March 1995, former First Lady Nancy Reagan delivered an unfavorable congressional speech asking, “Why is it we no longer hear the drumbeat of condemnation against drugs coming from our leaders and our culture?”77 Not losing popular support over the failed anti-drug policy was so important to Clinton’s reelection that the president used executive privilege to bar the public release of an FBI memo that was critical of the administration’s performance.78 This need to prevent the further loss of approval is consistent with the avoiding responsibility explanation for deference.
Second, McCaffrey’s appointment coincided with a change in the administration’s anti-drug policy, which is also consistent with avoiding responsibility. Initially, when Clinton ran for his first presidential term, he did not have a strong anti-drug policy agenda. In the presidential debates in 1992, he mentioned drugs four times — three times in relation to how his brother suffered from drug addiction, and once as part of a list of general problems pertaining to racial divisions in the United States.79 After becoming president, Clinton elevated the drug czar to a cabinet-level position, increasing the responsibilities associated with this job.80 He also cut funds for interdiction and source-country initiatives, focusing on treatment programs instead.81 When this approach did not yield sufficient results, the administration had to reintroduce more costly international anti-drug efforts, which included intensifying the use of military force in the source countries.82 The appointment of a four-star general as drug czar could be helpful in distancing the domestically focused president from these costly international moves.
Finally, when questioned in public about the effectiveness of the anti-drug policies, Clinton tried to minimize his responsibility. For example, during the presidential debates on Oct. 6, 1996, he was directly asked whether he bore some responsibility for the rise in drug use among teenagers. In response, he said: “Let me tell you what I tried to do about it. I appointed a four-star general who led our efforts south of the border to keep drugs from coming into the country as our nation’s drug czar.”83 After listing several other anti-drug initiatives, Clinton claimed that “we all” bore some responsibility and concluded by saying that the issue of drugs “shouldn’t be politicized.”84 During Clinton’s second term, the media criticized his administration’s policy of offering financial incentives to television networks for embedding anti-drug messages into their programming. In response, Clinton said that he thought that McCaffrey might have reached a conclusion that doing so would be a good thing. He then firmly asserted: “This was his [McCaffrey’s] initiative.”85 Thus, the president was making sure to distance himself from the seemingly unsolvable drug crisis. As McCaffrey put it, “Whenever we have an intractable problem, we tend to dredge up a czar and they fail.”86
As with trying to boost approval, when an effort to avoid responsibility drives the executive, the erosion of civilian control is a byproduct of civilian deference, not its goal. Nevertheless, the urge to avoid responsibility not only expands the military’s influence in politics and limits civilian input in policymaking, but it can also undermine the trust between the government and the armed forces if the former blames the latter when a policy fails.
Cajoling the Military
The third driver of civilian deference to the military that leads to erosion of civilian control is cajoling the military with power. When this takes place, the government delegates some policymaking tasks to members of the military profession to abate civil-military tensions. It is not only civilian officials who defer to senior military officers (both retired and active-duty). Fellow officers and soldiers also view generals and admirals with high levels of respect and veneration.87 Therefore, involving high-ranking members of the military profession in policymaking could be beneficial for curbing the armed forces’ opposition to the executive or its policies and decreasing civil-military tensions. Andrew Bacevich calls such bargaining the “dirty little secret” of American civil-military relations — the necessity of the executive to cajole and appease the military.88
The civilian government might also engage in power-sharing when it understands the “best military advice” as an ultimatum.89 This power dynamic is by no means unique to the United States. It was common, for example, in Brazil under Presidents Fernando Henrique Cardoso and Luiz Inácio Lula da Silva in the late 1990s and the first decade of the 2000s.90 Although the United States has never been under military rule (unlike Brazil), American military elites can exercise tangible political pressure. Resisting pressure from the military might be especially challenging for presidents without previous military experience. Indeed, the civil-military tensions of the Clinton era, or when Gen. Stanley McChrystal pressured President Barack Obama to increase the number of troops in Afghanistan, provide vivid examples of how the military’s advice can threaten civilian authority.91 To avoid a public standoff with the military over policy disagreements, the civilian government may yield power to the military to prevent or alleviate tensions. In addition, the possibility that opposing civilian elites (e.g., the opposition party in the legislature, or a potential presidential candidate) might exploit the military’s tensions with the government to their advantage could also compel the executive to appease the armed forces through deference.92
When cajoling the military with power drives civilian deference, the military’s dissatisfaction with policymaking will likely surface before power is delegated. This driver of deference is different from boosting approval and avoiding responsibility because in this case the executive’s decision is focused on the military. It is also different from avoiding responsibility, since in the case of cajoling with power, the government would strive to portray the relationships with the military as supportive and cooperative. However, if the government is trying to secure both popular support and the military’s support, cajoling the military with power can subsequently overlap with the other two drivers of deference. For instance, after appeasing the military by appointing a member of the military profession to a key policymaking position, the government may use the military’s involvement in politics to affect the public’s perception of the policy.
For example, over the eight years of the Obama administration, security policy underwent scrupulous civilian scrutiny, including multiple rounds of questions and edits.93 A particular concern was the number of civilian casualties in counter-terrorism raids in Yemen, Somalia, Syria, and Iraq. In contrast to his predecessor, Trump decided to relax civilian control of U.S. military operations, strikes, and bombings of terrorist targets in these states. In the first months of his administration, he proclaimed parts of Yemen and Somalia to be zones of active hostilities and expanded the authorities of U. S. Africa Command commanders to launch operations without high-level deliberations at the National Security Council’s committees.94 While contributing to the speed of operations on the ground, this policy change also decreased civilian input in security policymaking compared to the previous eight years. It therefore constituted erosion of civilian control by deference.
Why did Trump decide to loosen civilian scrutiny over counter-terrorism raids? To begin with, Obama-era interagency deliberations on defense and security were often criticized as micromanagement that stymied the policy process and slowed-down the operations.95 Colin Kahl, the former national security adviser to Vice President Joe Biden, admitted that the military requested more authority over counter-terrorism raids at the end of the Obama administration but did not receive it.96 Overall, the military officials did not approve of the Obama administration’s hands-on approach and repeatedly complained about it.97 Opinion surveys also indicate a low opinion of Obama as commander-in-chief among the military.98 For Trump, maintaining this high standard of civilian oversight would have led to undesirable civil-military tensions. Given his low and decreasing popular approval, he would not have been able to sustain as high a level of civil-military conflict as the previous administration. In addition, Trump’s reputation as a draft-avoider made him even more vulnerable to the military’s criticism.99 Thus, according to the second theoretical assumption of this study, delegating power to the military was more beneficial for the government than trying to keep it. These conditions motivated cajoling the military with power.
When cajoling the military with power drives civilian deference, the military’s dissatisfaction with policymaking will likely surface before power is delegated.
In addition, intensifying counter-terrorism operations in Yemen, Somalia, Syria, and Iraq were in line with Trump’s campaign promise to defeat the Islamic State, particularly through bombing campaigns.100 The delegation of authority to the military in January 2017 is consistent with Trump’s statement on June 2016 as a presidential candidate: “We have generals that feel we can win this thing so fast and so strong, but we have to be furious for a short period of time, and we’re not doing it!”101 Since Trump’s decision implied that civilians would be withdrawing from policymaking in this area, this example of deference is a better match for cajoling with power than the approval booster explanation.
It is crucial to keep in mind that cajoling the military can sometimes come into conflict with popular approval. When that occurs, executives can adjust their preferences and shield themselves from political criticism by underscoring the military’s responsibility for the decision. For example, Trump was fast to blame “the generals” within the Defense Department when a counter-terrorism raid in Yemen resulted in the first U.S. casualties.102 This was an attempt to avoid responsibility for the loss of life associated with the policy after the fact. However, his initial praise for “the generals” and the policy itself indicate that the root driver of deference in this case was cajoling the military.
The three drivers of deference are rooted in the triangular power relationship between the public, the government, and the military. Each stems from different intentions on the side of the government. Boosting approval is aimed at gaining more popular approval for government policies. Avoiding responsibility is supposed to prevent political losses by shielding the executive from criticism from political opponents, the media, or the public. Cajoling with power is meant to influence the military and reduce civil-military tensions. Depending on which of these three drivers of deference is at work, an executive’s policy preferences and participation in politics will differ (See Table 1).
Table 1: Drivers of Deference
| Boosting Approval | Avoiding Responsibility | Cajoling with Power | |
| Executive's Motivation | Increasing popular approval for policy | Not losing popular approval | Avoiding/decreasing tensions with the military |
| Executive's Policy Preferences | Strong, consistent with previous program/statements | Inconsistent with previous policies or promises | Overlap or at least do not undermine previous program/statements |
| Executive's participation in the policy process | Visible, joint with members of the military profession | Visible withdrawal/distancing from policymaking | De facto withdrawal/distancing, the projected image can vary |
Driven by these three motivations, civilian deference to the military launches the process of erosion of civilian control that can have devastating consequences for civil-military relations and broader political institutions and processes. Below I provide evidence of some of the destructive effects that Trump’s reliance on the military in politics had for the triangular power relations between the civilian population, executive officials, and the military.
Evidence of Erosion of Civilian Control
The executive’s deference to the military in pursuit of boosting approval, avoiding responsibility, or cajoling the military can have the devastating side effect of eroding civilian control. Below I use quantitative and qualitative data to show that, under the Trump administration, deference to members of the military profession increased the relative power of the military vis-à-vis civilians and thus led to the erosion of civilian control. Specifically, the evidence demonstrates that delegating policymaking tasks to the military decreased civilian input in policymaking, elevated the ability of elite military officers to impose constraints on policymaking, and called into question the armed forces’ non-partisan nature, disrupting the persistent trends of popular confidence in the military.
Decreased Civilian Input in Policymaking
To begin, appointing Mattis as secretary of defense marginalized civilians’ voices in national security policymaking. Mattis’ speechwriter, Com. Guy Snodgrass, writes in his memoir that to speed up the Defense Department’s work, Mattis preferred to surround himself with employees with a military background. Mattis specifically preferred those who had previously served with him at U.S. Central Command and Navy and Marine Corps officers. At the same time, unable to adapt to the military-dominated working style, a number of career civilian staffers left their jobs.103 As a result, the majority of the department-wide coordination meeting attendees were in uniform, including senior military assistant to the secretary of defense, Adm. Craig Faller, who led those meetings.104 The dominance of the military in the Pentagon at that time became so salient that the employees jokingly called it “U.S. Central Command—North” or “PENTACOM,” short for Pentagon Command.105
Not only did the number of uniformed personnel in the Pentagon increase, but so did the political power of some military members. During Mattis’ tenure in the Department of Defense, his chief of staff, Rear Adm. (Ret.) Kevin Sweeney, together with Faller, acquired enormous influence in day-to-day operations and communications in the department. This resulted in Faller, who was an active-duty admiral, issuing directives to senior civilian appointees who had been confirmed by the Senate.106 Snodgrass also reports that other uniformed members of Mattis’ team often had to assume an awkward position “when asked to go and ‘square away’ a senior civilian leader.”107 Thus, due to the delegation of policymaking responsibilities to members of the military profession, the military’s power over policy issues increased substantially, indicating an erosion of civilian control.
This change did not go unnoticed. By 2018, the president’s increased reliance on active-duty and retired members of the military became so salient that a National Defense Strategy Commission report included a warning that decreasing civilian input in defense and national security decision-making undermines civilian control of the military.108 Nevertheless, it is important to note that immediate observers of the militarization of the Department of Defense claimed that there was no malign intent on the part of Mattis and his team.109 Instead, they believed that to keep the Pentagon’s enormous machine moving, the team had to speak the same language.110 However, since this common language happened to be military parlance, the initial appointment of a member of the military profession in the key executive position led to the decrease of civilian input in policymaking and the further weakening of civilian control.
The civil-military balance of power only worsened throughout the Trump presidency. With many civilians leaving the Pentagon and about 40 percent of political appointments remaining unfilled by the administration, the relative power of the Office of the Secretary of Defense decreased while the influence of the Joint Staff within the Department of Defense increased substantially.111
The Military’s Political Weight in Domestic Politics Increases
Using the military as an approval booster and cajoling the military with power inflates the military elites’ sway in domestic politics and increases their ability to impose constraints on policy. Trump’s persistent delegation of policymaking prerogatives to members of the military increased the influence of senior military officers in policy debates.112 For instance, in June 2020 Trump considered invoking the Insurrection Act and using the military to disperse U.S. citizens participating in Black Lives Matter protests.113 In response, Mattis broke his long-time silence to denounce this move. In his letter, Mattis used his military background to question the president’s decision and criticize military leadership, saying:
When I joined the military, some 50 years ago, I swore an oath to support and defend the Constitution. Never did I dream that troops taking that same oath would be ordered under any circumstance to violate the Constitutional rights of their fellow citizens — much less to provide a bizarre photo op for the elected commander-in-chief, with military leadership standing alongside.114
The letter’s central argument was that it is unacceptable to have a militarized response to domestic popular protest and to politicize the military. In a similar manner, Mullen and Dempsey also publicly criticized the president’s intention to use the military for suppressing domestic protests.115 Paradoxically, the fact that these high-ranking members of the military profession felt that it was appropriate to use their military credibility in this way shows that Trump’s increased reliance on the military to improve public approval made the military elites powerful political actors capable of limiting his policy decisions. Moreover, while arguing against the militarization of politics and the politicization of the military, these high-ranking officers’ statements demonstrate that the military had already become a potent agent in domestic policymaking. The media coverage of these statements underscored this point: A Factiva search for “Mattis AND Trump AND protests” on the day of Mattis’ statement and the following day yielded 436 hits. A similar search for former Defense Secretary Robert Gates’ statement turned up only 14 matches. This shows how much preference the members of the military profession receive in media coverage and public discussion in comparison to prominent civilian politicians.
Using the military as an approval booster and cajoling the military with power inflates the military elites’ sway in domestic politics and increases their ability to impose constraints on policy.
Of course, this was not the first time in America’s recent history when military elites publicly criticized civilian leadership’s policies. For instance, when Clinton decided to abolish the ban on openly homosexual individuals serving in the armed forces, he became the target of harsh criticism from the military, including both active-duty and retired officers.116 During the Bush (43) administration, recently retired generals made public statements in opposition to how Secretary of Defense Donald Rumsfeld was managing the war in Iraq.117 McChrystal’s criticism of the Obama administration in a Rolling Stone profile cost him his military career.118 However, in these examples, the statements from military officers were not with regard to domestic politics in the United States. It was Trump’s increased reliance on the military in domestic political issues that brought the military elites to the discussion of police brutality, systemic racism, and the appropriate response to a protest in the streets of an American city.119
Democrats Trust the Military Less than Independents
As discussed earlier, using the military to increase public approval can undermine popular trust in the military as a non-partisan institution.120 This, in turn, could make popular confidence in the armed forces conditional on attitudes toward the president and his or her party. In the United States, Republicans historically have had higher confidence in the military.121 Therefore, analyzing whether the Trump administration increased the partisan divide concerning confidence in the military, it is important to keep this baseline in mind.
To see how the Trump administration’s early years affected the relationship between partisanship and trust in the military, I analyzed the data from the General Social Survey.122 Using logistic regression with year-fixed effect, I compared how party affiliation affected the likelihood of having high confidence in the military in 2018 in comparison to 2016. To account for potential confounders, I controlled for gender, race, education, age, and income.123 The findings suggest that in 2016 the likelihood of a Democrat having high confidence in the military was not significantly different from that of an independent. However, the results from 2018 show that two years into the Trump administration being a Democrat was associated with a significantly lower likelihood of having high confidence in the military than for an independent. With the long-term trend of high and increasing confidence in the military among Republicans staying intact, the partisan gap between how Democrats and Republicans feel about the military widened. In other words, the chances of having high confidence in the military had become a more partisan issue than in previous years.
Previous research indicates that support for the military can be a function of which party’s candidate is in the White House.124 To test this explanation, I conducted similar tests for partisan confidence in the military in the first two years of the Obama and Bush (43) administrations.125 The results suggest that during the first years of the two previous administrations, Democrats’ high confidence in the military increased as much as independents’ confidence. However, the results of the quantitative analysis from 2016–2018 show the disruption of this trend in the first years of Trump’s presidency. Unlike in 2016, being a Democrat in 2018 significantly decreased the chances of having high confidence in the military. This finding suggests that Trump’s extensive delegation of policymaking prerogatives to members of the military profession undermined the long-term patterns of confidence in the military among democrats. Future research will show whether this effect will persist or be reversed under the Biden administration.
Conclusion
This paper introduced the novel concept of erosion of civilian control by deferring to members of the military profession in policymaking. This concept systematically captures previously understudied instances of weakening civilian control that occur due to elected officials’ behavior and not a power-hungry or insubordinate military.
This study investigated three drivers that motivate civilian deference to the military based on the power dynamics between the government, the military, and society. It suggests that civilian executives defer to the military in policymaking to increase popular approval for their policies, avoid taking responsibility for difficult political decisions, and cajole the military with power. Trump’s delegation of policymaking tasks to the military showed the analytical utility of the new conceptualization of erosion by deference and the three drivers of this phenomenon. Additional examples beyond the Trump administration helped demonstrate that deference is not unique to one particular presidency and cannot be explained by Trump’s personal peculiarities. The subsequent analysis of quantitative and qualitative evidence shed light on how Trump’s deference to the military decreased civilian input in policymaking, enabled the military to impose constraints on policy, and made popular confidence in the armed forces more partisan than it previously had been. While the latter development is a warning sign of a partisan cleavage in U.S. civil-military relations, the former two are clear indicators of the erosion of civilian control by deference to the military in policymaking.
Understanding the nature and motivations of delegating policymaking prerogatives to members of the military profession has important implications for theory and policy. With regard to theory, it expands our understanding of the erosion of civilian control as a phenomenon that is not limited to the ambitions or misbehavior of the armed forces. Also, identifying the drivers of erosion by deference allows us to identify the conditions that are most conducive to this phenomenon taking place. On a more general level, having a nuanced perspective of the erosion of civilian control by deference contributes to an understanding of the mechanisms of democratic backsliding and the role that elected officials play in it.
From the policy perspective, acknowledging that erosion of civilian control by deference is a byproduct of elected officials navigating the triangular power relations can help civilian leaders to evaluate the costs and benefits of relying on members of the military in policymaking. Specifically, knowing that using the military to increase approval could make high-ranking officers powerful opponents of the government should cause policymakers to reconsider the benefits of this move. Similarly, understanding that an executive’s over-reliance on the military undermines popular confidence in the armed forces as a non-partisan institution should be a warning sign to the political institutions — first and foremost, the legislature — that provide checks on the executive’s decisions.126 In addition, public awareness of these drivers and the consequences of the executive’s deference to the military would help decrease the benefits and increase the costs of using the armed forces to increase approval and avoid responsibility.
The phenomenon of erosion by deference and its drivers are not unique to the United States. The conditions under which they occur are present in other democracies. In particular, high confidence in the military — an enabling factor of deference for increasing approval — is a widespread phenomenon in many democracies, including France, the United Kingdom, and Italy.127 Moreover, the rise of populist leaders who make unrealistic promises during electoral campaigns creates favorable conditions for erosion by deference driven by attempts to avoid responsibility.128 In addition, these new populist leaders, who come from outside the “system” and often lack policy experience, may be prone to tensions with the military elites, inviting erosion by deference for the sake of cajoling the military. Thus, having a clear conceptualization of the drivers of erosion by deference opens avenues for further investigation and preventing the weakening of civilian control in democracies around the globe.
Polina Beliakova is a Ph.D. candidate at the Fletcher School and assistant director at the Center for Strategic Studies at Tufts University. Her research focuses on civil-military relations and erosion of civilian control in democracies. Her current book project examines how governmental policies about the use of force can weaken civilian control.
Acknowledgements: The author is grateful to Monica Toft, Jim Golby, Roger Petersen; the participants of the Research and Policy Seminar at the Center for Strategic Studies, especially Thomas Cavanna, Sidita Kushi, Benjamin Denison, Nils Hägerdal, and Karim Elkady; and two anonymous reviewers for their helpful feedback on earlier versions of this manuscript.
Image: Department of Defense; Photo by U.S. Army Sgt. James K. McCann
Appendix 1
Table A: Effect of party identification on high confidence in the military in 2016–2018
| DV: high confidence in the military | Odds ratio | SE |
|---|---|---|
| Democrat = 1 | 1.087 | (0.136) |
| Year 2018 = 1 | 1.655*** | (0.170) |
| Democrat* Year 2018 | 0.659*** | (0.0969) |
| Republican | 1.616*** | (0.187) |
| Other party | 0.831 | (0.169) |
| Female | 0.740*** | (0.0545) |
| White | 1.122 | (0.0973) |
| Higher education | 0.795*** | (0.0631) |
| Ln (Age) | 1.253** | (0.121) |
| Annual household income: middle 1/3 | 1.052 | (0.0961) |
| Annual household income: upper 1/3 | 1.128 | (0.111) |
| Constant | 0.450** | (0.168) |
| Observations | 3,187 | |
| Psuedo R-squared | 0.0297 |
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Table B: Effect of party affiliation on high confidence in the military in 2008–2010
| DV: high confidence in the military | Odds ratio | SE |
|---|---|---|
| Republican = 1 | 2.327*** | (0.377) |
| Year 2010 = 1 | 1.148 | (0.116) |
| Republican * Year 2010 | 0.993 | (0.181) |
| Democrat | 0.990 | (0.120) |
| Other party | 1.520 | (0.474) |
| Female | 0.711*** | (0.0603) |
| White | 1.277** | (0.134) |
| Higher education | 0.695*** | (0.0637) |
| Ln (Age) | 0.872 | (0.0985) |
| Annual household income: middle 1/3 | 1.123 | (0.122) |
| Annual household income: upper 1/3 | 1.072 | (0.120) |
| Constant | 1.419 | (0.617) |
| Observations | 2,415 | |
| Pseudo R-squared | 0.0424 |
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Table C: Effect of party affiliation on high confidence in the military in 2000–2002
| DV: high confidence in the military | Odds ratio | SE |
|---|---|---|
| Democrat = 1 | 1.318** | (0.181) |
| Year 2002 = 1 | 2.036*** | (0.244) |
| Democrat * Year 2002 | 0.875 | (0.154) |
| Republican | 2.057*** | (0.259) |
| Other party | 0.716 | (0.245) |
| Female | 0.773*** | (0.0656) |
| White | 1.183 | (0.132) |
| Higher education | 0.713*** | (0.0683) |
| Ln (Age) | 0.935 | (0.108) |
| Annual household income: middle 1/3 | 1.160 | (0.124) |
| Annual household income: upper 1/3 | 1.083 | (0.116) |
| Constant | 0.616 | (0.270) |
| Observations | 2,460 | |
| Pseudo R-squared | 0.0398 |
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1