The Trump Self-Coup Attempt: Comparisons and Civil–Military Relations

Abstract

The storming of the US Capitol building in January 2021 was a presidential attempt at a self-coup. To make the case, this article reviews elements of the Capitol assault and the events leading up to it, in light of the key conceptual components of a self-coup, and how those compare to attributes of other kinds of attacks on governments. The Trump self-coup will then be compared and contrasted empirically to other self-coups perpetrated by leaders. It is found that what separates successful self-coups from those that fail is whether the military backs the undertaking. Thus, a section is included on US military behaviour in response to Trump's attempts to gain military adherence for his political actions.

The extraordinary attack on the US Capitol on 6 January 2021 caught many by surprise. The Congress had not suffered such an unfortunate fate since the war of 1812 when the British burned the Capitol building to the ground. Journalists, scholars and politicians quickly attempted to characterize the event, calling it variously an insurrection, a mob invasion, a seditious act, a legitimate protest, a rebellion or a coup. This article argues that the attack had all the telling features of a self-coup attempt. Self-coups occur when a nation's chief executive, in order to hold onto, consolidate or expand power, coercively interferes with or shuts down another branch or branches of government. In his desperate attempt to stop the certification of the election of Joe Biden and Kamala Harris, President Donald Trump helped instigate his followers to storm the Congress in order to stop the electoral count underway.

This attempt, which ultimately failed, was the culmination of a longer-term effort by President Trump to cast aspersions on the democracy by encouraging his followers – many of whom were violent extremists – to attack electoral institutions and personnel charged with counting and certifying votes. These trends will be analysed. Then the Trump self-coup will be compared to other kinds of violent and non-violent assaults on government, identifying the key traits that distinguish the self-coup from other actions. It will then compare and contrast Trump's actions with self-coups launched by other chief executives from nominally democratic systems. As we find that what separates successful self-coups from those that fail is whether the military backs the undertaking, we also include a section on the behaviour of the US military in response to Trump's attempts to gain military adherence for his political actions. This is followed by a brief analysis of how the Trump self-coup could have longer-term, deleterious impacts on the democracy. We conclude with a discussion about the rightful place of military dissent in the face of presidential decisions that may harm democracies.

What happened on 6 January 2021? What preceded it?

Thousands of unruly citizens stormed the Capitol building on 6 January, pushing past Capitol police, causing property damage and threatening harm to congresspeople and the vice president. This was not a spontaneous gathering of protesters. It was an organized effort to stop the Joint Session of Congress from certifying the election of Joe Biden and Kamala Harris as president and vice president. And it was an act that had been encouraged, indeed instigated, by the president of the US in the months, days and moments leading up to the invasion (Rutenberg et al. 2021).

Participants included members of right-wing militias, terrorist groups, neo-Nazi and white supremacist organizations, along with followers of conspiratorial groups such as QAnon. Many came armed. These groups pre-dated the Trump presidency and certainly had agendas of their own. The issue is whether the president exhorted them to take the measures they did on 6 January as a means of furthering his own goals of overturning the election results of 3 November by obstructing congressional deliberations. And did these groups and individuals understand his intent, and respond directly to his exhortations? There is strong evidence that they did just that, bolstering the contention that this was a self-coup attempt.

Well before 6 January, and indeed well before the election of 3 November 2020, President Trump had set the stage for the self-coup by encouraging the very groups that would violently storm the Capitol. This dates as far back as 2017, when he noted there were ‘very fine people’ among the neo-Nazis, white supremacists and other racist groups that stoked violence at a rally in Charlottesville, Virginia, protesting the proposed removal of Confederate monuments. White supremacist leaders reacted by thanking the president for his remarks. This was part of a larger pattern whereby the president seemed to legitimize violent right-wing organizations. In what was his most direct endorsement of domestic extremism, President Trump, in his first debate with Joe Biden on 29 September 2020, not only refused to condemn white supremacists, but exhorted them, saying, ‘Stand back and stand by.’ The Proud Boys and other extremist groups interpreted those remarks as a licence to utilize whatever means were at their disposal to carry out the president's plans (Gabbatt 2020).

But what motivation would these groups and countless millions of his individual followers have for rallying to Trump's defence to subvert the democratic process? The president framed the issue as one of electoral fraud; the voting mechanisms could not be trusted. Despite courts having ruled against this contention dozens of times, finding that such charges were unsubstantiated, the president persisted with his claims that the election had been rigged, right up to the attacks of 6 January, and beyond. These baseless claims stirred up those of his followers who felt aggrieved. At a rally more than a month before voters went to the polls, Trump alluded to the possibility that the electoral count would be disputed (Washington Post 2021: 10). After the election, throughout the autumn and winter, Trump leaned on election officials in states such as Georgia and Arizona with a blizzard of tweets and personal phone calls, trying to get them to undo the results of the election. When that failed, he turned his focus to 6 January, when Congress would certify the election results by counting the electoral college votes from each state, historically a proforma ritual. Behind the scenes, the president zeroed in on three manoeuvres in his attempt to overturn the election. He pressed Justice Department officials to assert there were irregularities in the vote. He goaded state officials to reopen the counts. And, as a last resort, he kept lobbying his vice president simply to cast aside the results on 6 January (Washington Post 2021: 56).

Trump used social media such as Twitter to propagate false narratives and to urge followers to come to Washington DC on the 6th ‘to stop the steal’. ‘Big protest on January 6th’, he tweeted on 19 December. ‘Be there, will be wild’ (quoted in Barry and Frenkel 2021). Trump could have anticipated at that time that his invitation would result in a violent assault on the Congress, since his stated objective was to stop the electoral count, and since he knew full well that among his supporters were thousands representing the most dangerous and armed extremist groups in the country (Feuer 2021a). He could have surmised that members of such organizations would interpret his words in ways that could result in violence. Indeed, rioters who were apprehended have since acknowledged that they were acting on the instructions of President Trump (Mangan 2021).

This was not a spontaneous outpouring of wrath directed at the Congress. Those groups responded by spreading the word and coordinating their planned attack through social media platforms. In communications weeks before, they made no secret of their desire to take arms into the District of Columbia (Feuer 2021b; Washington Post 2021). Weeks prior to the attack, and well before being apprehended and charged as defendants in criminal cases, participants in the mob attack had already indicated they were going to Washington DC at the behest of the president, not for a peaceful rally, but for an invasion of the Capitol (Feuer 2021b; Kevin Johnson 2021).

As for the president, his remarks indicate that he fully intended to disrupt the congressional proceedings in order to stop the transfer of power to a new president. In what can be taken as a veiled threat, Trump tweeted on the day before the attack, ‘I hope the Democrats, and even more importantly, the weak and ineffective RINO [Republicans in name only] section of the Republican Party, are looking at the thousands of people pouring into D.C. They won't stand for a landslide election victory to be stolen’ (quoted in Frum 2021).

At a planned rally on the day of the attack, the president delivered an incendiary speech, exhorting his followers to ‘stop the steal’, adding, ‘We will never concede. It will never happen. You don't concede when there is theft involved.’ Repeatedly, he said there was a need to fight. He vilified weak Republicans, and then said, ‘You have to get your people to fight.’ Then he told the crowd to march down Pennsylvania Avenue to the Capitol, where the Congress was in joint session. Trump said, ‘After this we're going to walk down, and I'll be there with you.’ ‘You have to show strength, and you have to be strong,’ he added (quoted in Sherman 2021).

Evidence suggests that those at the rally clearly understood the president to mean they should storm the Capitol to disrupt the election count proceedings. At the rally many were heard screaming, ‘We're storming the Capitol!’ (Mogelson 2021). In its impeachment resolution, the House of Representatives directly accused the president of inciting a mob to act violently and illegally to prevent the joint session from certifying the election results (US House of Representatives 2021).

The 6 January attack on the Capitol, which occurred outside the bounds of law and was violent in nature, was planned and instigated by the president for the purpose of preventing the Congress from ratifying the election of Joe Biden as president. It had all the hallmarks of a self-coup (Hill 2021).

What is a self-coup? How does it differ from related concepts?

A self-coup is an effort launched by a nation's chief executive to hold onto, consolidate or expand power by interfering with or shutting down another branch or branches of government (Cameron 1998; Cline Center 2021; Koonings 2021; Svolik 2014).¹ If a self-coup has been successfully executed, then a president will be able to rule without the constraints another branch of government can impose. In successful self-coups, presidents hold onto or expand power at the expense of legislative and/or judicial power.

The notion of a self-coup can get quickly lost in an extended family of related concepts, with a classic military coup being the obvious first comparison (Marsteintredet and Malamud 2020). But there are numerous kinds of attack on government beyond military coups that share some but not all traits with self-coups. These include other executive-led manoeuvres, aggressive actions by other branches of government, and civilian-led attacks on the state by outsiders. To help sort out the differences, we enumerate some key traits in Table 1 and identify which kinds of governmental assaults assume which characteristics. All attacks on government can be distinguished by their motives, who the perpetrators and targets are, what are the methods used and what outcomes are derived.

Table 1 Self-Coup-Attempts: Attributes and Comparisons with Related Concepts

Notes: ^aCore attributes of self-coup attempts:

A Organized, non-spontaneous action

B President initiates

C Illegal

D Coercion, or threat of use

E Target: non-executive branch of government

F Purpose: suspend or impede non-executive branch

G Purpose: expand executive's power and/or extend rule

H Outcome: fails to suspend or impede non-executive branch

I Outcome: fails to expand executive power or extend rule

Y, Yes; M, Maybe; N, No; N/A, Not applicable.

The self-coup shares some features with a military coup: they are organized, illegal and coercive. What separates them first and foremost is the fact that self-coups are always initiated by the president or prime minister him- or herself. The military or some other state security force might play a role in supporting and/or enforcing a self-coup, but they do not initiate it. Coups are, according to Jonathan Powell and Clayton Thyne (2021), ‘illegal and overt attempts by the military or other elites within the state apparatus to unseat the sitting executive’. Hence, coups are almost always military led, but also aimed at a different target: the chief executive him- or herself, with the intention of dislodging the president or prime minister from office, and then replacing him or her with either temporary caretakers or long-term rulers, often in the form of juntas. Having seized executive office, militaries may continue their assaults on other branches of government. However, there are numerous examples of military regimes that allow congresses and courts to continue functioning (Gandhi 2008).

While the armed forces can single-handedly bring down a government, presidents need co-conspirators, since they may have no coercive means at their disposal, and since their actions are not within the bounds of constitutional authority and therefore will be neither legally enforceable nor voluntarily complied with by congresses, parliaments or courts. Self-coup implementers include the armed forces, other state security forces, political allies (including complicit party leaders, legislators or judges) and non-state actors, such as civil society groups and individuals. However, evidence (provided below) strongly suggests that, without military backing, self-coups will fail, making the armed forces the most valued of co-conspirators.

Insurrections, rebellions and related forms of political action (e.g. revolutions) are aimed at some established authority and are always organized and launched by civilians from outside the government, not from within. They too share with self-coups the features of being organized, illegal and violent. Their intent is to pose a challenge to an established power by attacking, obstructing or overturning that authority, and they may or may not be directed at a non-executive branch of government.² In extreme forms, such as revolutions, their goal is nothing short of regime change.

Though the Trump self-coup depended on an assault by civilian outsiders against the Congress, his direct role in instigating the assault differentiates that episode from the others. It should also be mentioned that insurrections which take on non-violent forms are usually classified as acts of civil resistance. Civil resistance is a ‘form of collective action that seeks to affect the political, social or economic status quo without using violence or the threat of violence against people to do so’ (Chenoweth 2021: 1, emphasis added). They share with insurrections the trait that they are coordinated, organized, non-institutional and disobedient, working outside of conventional political processes (such as voting, lobbying).

Self-coups are distinct from more incremental, often legal, episodes of executive overreach. Nancy Bermeo (2016) refers to executive aggrandizement when executives weaken checks on their power incrementally. Presidents have measures at their disposal to tilt power in their favour that are done slowly and legally, within democratic institutions and processes. They can, for example, cajole their own political party or majority coalition to pass legislation giving them more authority to chip away at democratic safeguards. They could exploit crises as excuses for overreach, resorting to the legal use of states of emergency, special decrees or executive orders to bypass the will of the legislature. They may call for a referendum to elect a constituent assembly charged with rewriting the constitution to further strengthen their unchallenged authority.

These kinds of executive manoeuvres are often associated with the broader phenomenon of democratic backsliding. A chief executive wins office electorally, and then with the help of a compliant legislature and party loyalists, chips away at democratic processes, laws and norms, assuring for him- or herself and their political party maximum power. Stephan Haggard and Robert Kaufman (2021: 1) define backsliding as follows: ‘The incremental erosion of democratic institutions, rules and norms that results from the actions of duly elected governments, typically driven by an autocratic ruler.’ While the results of backsliding measures such as executive overreach may set the stage for self-coups, or achieve the same or similar results, they are procedurally distinct, since they utilize legal, democratic means to advance what ultimately may be non-democratic ends.

Legislative coups target an incumbent president. They occur when normal legal procedures for removing a president – ones that often fall under the framework of an impeachment – have been circumvented. Procedural violations could include a failure to provide evidence at the congressional trial, an unreasonable acceleration of the process, or a failure to give a president adequate time to prepare and present a defence. How extensive such procedural violations have to be to warrant a coup classification is a judgement call. These proceedings must be distinguished from legislative actions that are conducted in coordination with, and as part of, a military coup operation (Pérez-Liñan 2007: 50–51).

Finally, distinctions have to be made between successful self-coups and self-coup attempts. Because the Trump attack was unsuccessful, it is categorized here as an attempt, and all other phenomena are compared to it. A successful self-coup occurs when the president is able to hold onto power by rendering powerless another branch of government or disrupting its proceedings. We define self-coup attempts as ones that are rebuffed rather quickly and where the president must relinquish power.

What counts as success or failure also depends on what presidential intentions are. Some leaders intend to land crushing blows on the legislature or courts, seeking their closure or dissolution. If successful, these plans lead to either a short-term disruption or long-term break in the constitutional order (Pérez-Liñan 2007: 48–54). Other self-coups may have less ambitious goals, seeking to thwart a legislative initiative or judicial action without closing down the institution and rupturing the political order. However, unlike a normal act of executive lobbying, overreach or circumvention, this would take the form of a coercive action or threat, producing a speedy outcome beneficial to the chief executive.

A presidentially directed self-coup is a bold, coercive and risky move, and one used infrequently (see below for data). This raises the question as to why executives would choose that strategy over alternatives. Clearly if democratic chief executives could gain what they wanted by forging deals with congress and parties of the opposition, or by resorting to legal decrees or executive orders, there would be little reason to launch an incumbent takeover – unless intentions were, from the outset, to move towards a thoroughly autocratic scheme of governance.

Gretchen Helmke (2017: 68) argues that the probability of self-coups increases when the stakes are high, and when congressional deals are foreclosed, or presidents mistakenly believe they are. Executives faced with a recalcitrant legislature – where their party is in the minority and unable to build winning coalitions with other parties – are especially incentivized to launch a self-coup attempt (Helmke 2017: 59). But in addition, she argues that presidents calculate what the risks are that they will be punished by the governing branch they are attacking. Courts can rule executive moves to be unconstitutional. Legislatures can retaliate through impeachment or motions of no confidence. If they do, then according to Helmke, it is the end of the line for the president or prime minister.

However, a president's fortunes can be enhanced considerably by the armed forces. What the military might do in defence of an executive self-coup attempt is never contemplated by Helmke. Any legal, retaliatory measures contemplated by another branch can be easily annulled when militaries choose to back up an incumbent's bid to alter the balance of governing power. Thus, the entire illegal, coercive nature of an incumbent takeover lies outside the boundaries of her analysis and those of others (e.g. Svolik 2014) who predicate all political calculations being undertaken in the confines of a fully functional constitutional legal system of compliance and enforcement.

The Trump self-coup in comparative perspective

The Cline Center for Advanced Social Research has compiled a data set of 943 coups and auto-coups (self-coups) from 1949 to 2019. Included is information on whether coups took place, were successful, the actors who initiated them, and the fates of leaders (Cline Center 2021). Based on their coding, and our recoding, a total of 31 self-coups have been identified.³ The Cline Center analysts did not differentiate countries according to regime types, something we have done using the Bjornskov–Rode regime data set (Bjornskov and Rode 2021). This data set allows us to differentiate between democracies – both presidential and parliamentary – and autocracies – both civilian and military types – as well as monarchies. As shown in Table 2, 63% of self-coups began under dictatorships, both civilian and military in nature; 27% occurred under democracies and 6% under monarchies. The most successful self-coups occurred under dictatorships (18 of 21, or 86%) while the least successful occurred under presidential democracies (2 of 7, or 28.6%). Only two self-coups took place during parliamentary rule, though both of those succeeded.

Table 2 Self-Coups and Regime Types: 1951–2021

Source: Bjornskov and Rode (2021).

Disaggregating this data, we were able to identify nine cases within minimally democratic systems (see Table 3). It made sense to compare the Trump self-coup only with those that had occurred under ostensibly democratic governance to rule out autocratic systems that had rubber-stamp congresses and/or courts. We wanted to compare executive attempts to disrupt or dissolve other branches of government that were viable, with independent bases of power and authority.

Table 3 Self-Coups in Comparison: The Military Impact

Can the US be compared with the other self-coup democracies? It is certainly an older and more consolidated democracy than the others shown in Table 3. But it is not unique, and is also one that has experienced a dramatic decline in the last decade, according to Freedom House (2021). This downward trend is attributable to discrimination, the influence of special interests and partisan polarization. These trends have accelerated under Trump (Freedom House 2021: 1), who bears a special responsibility for sharpening the partisan divide in ways which encouraged many of his followers to perceive opponents as implacable foes that had to be defeated at all costs. This in turn, we would argue, fuelled the invaders of 6 January to view their actions as an all-or-nothing struggle to ‘save the nation’ from its worst enemies. Hence the US democracy no longer has an esteemed status in the world, making comparisons with other cases justifiable. To compare and contrast self-coups within putatively democratic states, we will examine motives and objectives, targets, perpetrators and results, as summarized in Table 3.

All self-coup attempts were driven by legislative and electoral difficulties or defeats, or high court obstruction. In Guatemala, President Jorge Serrano was being investigated for corruption, which could, he feared, have led to calls for his impeachment (Cameron 1998). In Indonesia, impeachment proceedings were already underway against President Abdurrahman Wahid when he attempted a self-coup (Aglionby 2001). Other cases involved difficulties that chief executives had in pushing through desired legislation. In Peru, President Alberto Fujimori's party only enjoyed a handful of seats in the congress, which made legislative passage difficult, but not impossible, were he inclined to build coalitions of support – which he was not. When, in 1992, the president demanded approval of decrees giving him extraordinary executive powers to fight terrorism, congress resisted (Cameron 1998). That was all the president needed to trigger his self-coup, shutting the legislative branch down. In El Salvador, President Nayib Bukele accompanied soldiers and police into the chamber of the country's Legislative Assembly to disrupt its proceedings and issue a warning. He was frustrated that lawmakers had delayed approval of a loan to fund his domestic security plan (Pion-Berlin and Acacio 2020).

The Gambian chief executive, Yahya Jammeh, provoked a self-coup after having been defeated in elections. Jammeh, like Trump, claimed ‘there were serious and unacceptable abnormalities in the elections’, following his defeat by opponent Adama Barrow on 1 December 2016 (Ozer 2019: 200). While admitting some errors, the Gambian electoral commission found no widespread irregularities that would have altered the outcome. Nonetheless, Jammeh refused to concede.⁴ Likewise, Donald Trump refused to admit defeat in the face of an impressive victory by Joe Biden in the 2020 presidential elections. Without evidence, Trump claimed that the election had been stolen from him due to various forms of fraud and misconduct. This claim has been repeatedly debunked (Funke 2020).

In all cases, presidents aimed their attacks on the congressional and/or judicial branches. Some, including President Trump and Salvadoran President Bukele, wanted to disrupt legislative or court proceedings in order to effect a desired result. Other leaders went so far as to decree their dissolution (Pakistan, Peru, Uruguay, Indonesia).

What separates successful from failed self-coup attempts is whether presidents achieved their stated goals, which in turn hinges on the role of the armed forces. In the cases of Pakistan, Peru and Uruguay, the presidents declared their intent to dissolve the legislative branch, and did so. Those legislatures remained closed for years or were transformed into pliant institutions (Peru). In Ecuador (2005) President Gutierrez tried to retire the entire high court in March 2005, but that triggered a vote in congress a few weeks later to depose him. When he refused to leave office, the military withdrew their support for him and then escorted him out of the presidential palace (Washington Post 2005). The Supreme Court justices remained in place.

As was the case with all self-coup attempts that ultimately failed, the decisive factor was the withdrawal of military support for the initiative. As shown in Table 3, all self-coups (except for El Salvador) that failed had no military support. For instance, in Indonesia, Gambia and Ecuador, presidents declared states of emergency to justify closing down legislatures and courts, but the armed forces in those countries refused to back them. Following military dissent, each of those presidents was compelled to resign. All those that succeeded in shutting other branches of government down had the military on their side (Pakistan, Uruguay, Peru). In the case of Peru, the military left the barracks to encircle the congressional building with tanks, prohibiting lawmakers from entering to perform their duties.

The case of the El Salvador self-coup is a peculiar one that on balance must still be considered a failure. Yes, President Bukele received the backing of the armed forces in his attempt to coerce the congress into voting to approve a loan he wanted. He interrupted congressional proceedings, accompanied by 50 heavily armed soldiers and police, issuing an ultimatum giving legislators one week to approve the loan, or if not he would provoke a popular insurrection and dissolve the congress. Bukele's self-coup did not succeed in the end, because when the Supreme Court a few days later ruled that he had violated the constitution, the president withdrew his threat and the congress resumed its normal duties.

The US case of a failed self-coup and civil–military relations

In the US there is no evidence that the military coercively backed the violent attack on the Capitol building on 6 January 2021. To the contrary, the army – with some delays – authorized the National Guard to deploy to halt the invaders. If this had been a military-supported self-coup attempt, we would see evidence that the armed forces, including the National Guard, had fulfilled a presidential desire that they facilitate the invasion. That is not the case.

This is not to say that President Trump did not attempt to manipulate the military into supporting his political agenda. He closely identified and ingratiated himself with the armed forces in a bid to win support for all his initiatives (Bender 2021; Brooks et al. 2021; Golby and Feaver 2021; Joyner 2021; Leonnig and Rucker 2021; Schake 2021). Early in his term President Trump had two retired four-star USMC generals and one active duty three-star US general in his administration. He frequently referred to ‘his generals’, failing to understand that the armed forces were sworn to defend the constitution and not him personally. Jim Golby (2021) lists 46 what he terms ‘civil–military incidents’ during the Trump administration, of which he states, ‘There are, however, many almost indisputable cases where Trump used senior military leaders or other military personnel as “swords or shields” to attack his political opponents or to defend himself or his administration from criticism.’

At the same time, Trump was not averse to shielding himself from scrutiny by deflecting blame onto the military. He would oscillate between heaping praise on his generals and scorn, depending on whether commanders were sufficiently committed to furthering his political aims. Loyalty was Trump's litmus test for good military behaviour, and he had no hesitation in firing (or threatening to fire) the best and the brightest officers should they not do his bidding. But this carrot-and-stick strategy could not win him the military allegiance he needed to impede congressional proceedings on 6 January.

Despite President Trump's efforts to have military leaders uncritically align with him, he was clearly unsuccessful in at least two critical instances, and consequently unsuccessful overall. The first, in the lead-up to the presidential election on 3 November 2020, took place on 1 June 2020 in Lafayette Square across from the White House, during protests against police brutality. It was a catalytic event, making it evident that the president would later on seek the support of the US military for a self-coup. The second occurred immediately after the storming of the Capitol on 6 January 2021 and before the inauguration on 20 January of that year.

During the widespread protests over police brutality towards Black Americans, President Trump wanted military enforcement of government decisions on handling protesters. According to Kori Schake (2021), Trump considered invoking the Insurrection Act of 1807 to give him the authority to deploy active-duty troops as law enforcement. While that option was under consideration, the secretary of defence talked to governors about the need to ‘dominate the battlefield’, and both the secretary and the chairman of the Joint Chiefs of Staff, General Milley, walked with the president through a public space that had just been forcibly cleared of peaceful protesters, in what was essentially a photo opportunity for the president as he walked to church.

Subsequently both the secretary and the chairman apologized for their participation in Trump's photo opportunity. Observers agree that, for senior US officers, the Lafayette Square event was a wake-up call to consider the grave political implications of their behaviour. One of the principal results of the walk to the church for the photo-op was General Milley's realization that he had been used by Trump as a political prop (Leonnig and Rucker 2021). From that time on, Milley was on high alert to any hint by the president that he might use the military for political purposes. When Milley publicly apologized for his participation in the church photo-op, he did so during his National War College commencement speech, without prior knowledge or clearance from the White House. When Trump later lashed out at him in the White House, Milley told him,

Mr. President, this has nothing to do with you. This had to do with me and the uniform and not politicizing the uniform. I'm not apologizing for you. I was apologizing for me. Mister President, I don't expect you to get that. But I'm a soldier, and I can never allow the politicization of the uniform. I can't do it. It's wrong. And that's why I apologized. (quoted in Leonnig and Rucker 2021: 230)

When Trump threatened to use the Insurrection Act to force the military to become involved in domestic situations, the generals reminded him of the very high political price he would pay (Leonnig and Rucker 2021: 230). Michael Bender writes,

Milley started sending clear signals to the White House that, whatever Trump's plan, he should leave the military out of it. ‘We do not take an oath to a king or a queen, a tyrant or a dictator’, Milley said at the opening of the U.S. Army's Museum on November 1, as Christopher Miller, the acting defence secretary sat nearby. (Bender 2021: 358)

Clearly, then, from at least the Lafayette Square event on 1 June, the top echelons of the military were aware that President Trump had no compunction in utilizing the military for his political goals and were signalling their resistance. These political goals would eventually include the self-coup attempt that unfolded on 6 January 2021. In testimony before the House of Representatives, Christopher Miller said that the president told him to do whatever was necessary to protect the demonstrators who would be gathering in Washington DC, including using the National Guard. Miller believed that Trump and his advisers were contemplating a call for martial law, which could have justified military intervention (Wolfe 2021).⁵

While we are reasonably certain of the events of 6 January and those leading up to it, there are, not surprisingly, different interpretations of why it took as long as it did to bring in the National Guard to clear the Capitol of the mobs. According to Carol Leonnig and Philip Rucker (2021), after giving his speech, Trump returned to the White House to watch the spectacle on TV. He refused any requests to call off his supporters as they invaded the Capitol, nor did he respond to pleas to provide protection to members of Congress and the vice president. Washington officials requested National Guard assistance but could not get appropriate permissions from the Department of Defense or the White House with a president in absentia who was more concerned about whether Vice President Pence would actively turn the count of electoral votes into a win for him. Apparently, it was Trump's daughter, Ivanka, who finally convinced him that he needed to take action. If he did not, the Trump legacy would be tarnished forever (Leonnig and Rucker 2021).

Once Trump agreed to act, the appropriate permissions were granted, and Maryland and Virginia National Guard joined the Washington DC National Guard in eventually bringing the situation under control, permitting the counting of electoral votes and the certification of Biden as the winner of the 2020 election to go forward (Leonnig and Rucker 2021). Though the self-coup failed to stop Congress from conducting its business later that evening, delays clearly attributable to the president alone were costly, as the late arrival of National Guard units failed to stop the invasion, loss of life, physical destruction of property and the threats against congresspeople and the vice president once the mob had entered the chambers. In other words, the self-coup attempt did move forward, but did not succeed in the end.

Shortly after 6 January, the top officers of the Joint Staff and the six chiefs of the armed services (including the Space Force and the National Guard) issued a ‘Memorandum for the Joint Force’ committing the forces to defend the constitution and respect the results of the presidential election. Specifically, the memorandum says, ‘As Service Members, we must embody the values and ideals of the Nation. We support and defend the Constitution. Any act to disrupt the Constitutional process is not only against our traditions, values, and oath; it is against the law.’⁶

The memorandum (in the Online Appendix), though released on 12 January, is a good indicator of the military's thinking about civil–military relations leading up to the mob invasion. In short, despite Trump's extensive efforts to win US military support for his political ploys, he was unsuccessful. Military commanders here drew a sharp distinction between, on the one hand, their oath to defend the constitution and their duty to obey the law, and on the other hand the illegal, unconstitutional and violent actions of 6 January 2021 which Trump had instigated (Woodward and Costa 2021).⁷ The memo also reveals that the military high command would do nothing subsequent to the mob attack to thwart the inauguration of President Biden eight days later, on 20 January, stating that ‘President-elect Biden will be inaugurated’ (see Online Appendix). This was yet a further expression of military dissent from the president's wishes that Joe Biden not succeed him. Without military backing, the self-coup attempt (and all other efforts to overturn the results of the November 2020 presidential election) was destined to fail and did.⁸

The self-coup legacy

If the Trump self-coup attempt was a last-minute, desperate gambit to subvert democratic process in order to hold onto power, it was not the president's final act, nor that of his followers. The ex-president's attempts to cast doubt on the integrity of the US electoral system persist, and as of November 2021, a substantial majority (68%) of Republicans still believed the election was stolen from Trump, while 6% of Democrats and 26% of Independents held the same view (Public Religion Research Institute 2021). Party identification has much to do with this, but clearly suspicions about the integrity of the elections are also stimulated by unsubstantiated claims of electoral fraud within the conservative media, and by Trump himself who, a year after the election, was still claiming he had won.

Those suspicions have led to partisan audits in Arizona, which failed to reveal any evidence of voter fraud, nor could they overturn the results of that state's election, which went to Biden. Equally damaging to US democracy is the fact that huge numbers of Republican national legislators have refused to denounce Trump and others who continue to propagate the lie about the 2020 elections. This lie, as maverick Republican representative Liz Cheney put it, ‘is an attack on the core of [the] constitutional republic’ (Cheney 2021). When the party faithful at mass and elite levels cast aspersions on what are valid and secure electoral processes, they contribute to the degrading of the democracy. Thus, the self-coup's pernicious impact on the US democracy continues to be felt.

Conclusion

The attack on the US Capitol building on 6 January 2021 is evidence of a self-coup attempt. When analysing the key elements of an incumbent takeover, it is clear that the attack fits the definition closely. Other kinds of assaults on governing entities may share some characteristics with self-coups, but not all, permitting us to rule out the Trump-led assault as anything other than an attempted incumbent takeover. While unprecedented in US history, Trump's efforts are comparable to those taken by foreign chief executives in the past who were equally intent on undermining the proceedings of another branch of government to prolong their tenure and enhance their power. What separates the successful self-coups from those that failed is the support of the armed forces. The US armed forces never came to the president's defence in pursuing this perilous course of action, and ultimately intervened to stop it.

It could be argued that a military that dissents from a legitimately elected president's wishes is acting politically, contrary to its professional duty to remain subordinate to civilian control. The phrase ‘acting politically’ raises a crucial issue in the analysis of civil–military relations in the US and elsewhere. According to the ‘normal’ theory of US civil–military relations, the operative phrase is ‘civilians have a right to be wrong’ (Golby and Feaver 2021; Huntington 1957). This right allows for statesmen to prevail even as they make unwise decisions. If politicians are foolish, then the presumption is that they will be subject to the harsh judgement of voters at the next election. With the check of elections, there is no need for the armed forces to dissent.

The problem is that this right to be wrong can be abused and was under the Trump presidency. Well before voters cast their ballots, considerable damage can be done in the short to medium term, not only to national security, but to civil–military relations and democratic norms and standards as well. The military has the right to oppose executive commands that violate the law or urge conduct that is manifestly immoral (Milburn 2010). But the military can also dissent from orders that undermine democratic safeguards and individual rights and freedoms (Pion-Berlin and Ivey 2021). The military was thus within its right to take exception to the president by not supporting his self-coup attempt, one which – aside from being illegal – could have placed the democracy in real peril by invalidating the result of a legitimate election and preventing the transfer of power to a new president-elect. In short, military dissent is not tantamount to insubordination and can be reasonably predicated on numerous grounds, on the understanding that, while sometimes justifiable, it is an action that should be used infrequently and cautiously.

While the military should never engage in politics, nor be forced by a president to be deliberative, it has an obligation to be aware politically (Crowe 1993). It should assess the political implications of the president's actions as well as its own. It is clear from our analysis that, fortunately for the US, the US military leaders did become increasingly sensitive to the politics of Trump's commands, and increasingly wary of blindly following his lead.