Why NATO survived Trump : The neglected role of Secretary General Stoltenberg

NATO only just survived the Presidency of Donald Trump. Trump had distinguished himself from virtually all US Presidents since the Second World War in his active hostility toward the alliance during the presidential campaign. When in office, he repeatedly toyed with the idea of withdrawing from NATO and was on the verge of publicly doing so at the 2018 NATO summit. But whereas the President withdrew the US from the Iran Nuclear Deal, Paris Climate Agreement, UNESCO, and undermined the WTO, WHO, UN Refugee Agency, and the Green Climate Fund from within, Trump eventually changed his public position on NATO in 2019. In his state of the union speech in February, he described his tentative change of mind: ‘For years, the United States was being treated very unfairly by NATO — but now we have secured a $100 billion increase in defence spending from NATO allies’ and at the London Leaders meeting in December, he declared that ‘NATO serves a great purpose.’

While the dust has barely settled on the Trump Presidency, three types of explanations can be deduced from general analyses of Trumpian foreign policy. The first locate the sources of Trump's relatively continuous NATO policy on the domestic level. Some argue that the US foreign policy establishment constrained the Trump Administration's foreign policy impulses and ensured continued support for NATO. Others point to inherently expansionist tendencies of liberalism, allegedly entrenched in US society, that prevent a constrained foreign policy and withdrawing support for NATO. 4 The second camp emphasizes that continued support for NATO is the rational utility-maximising behaviour associated with US hegemony. 5 The third camp directs attention to Trump's idiosyncratic personality and cognitive features to explain his erratic and seemingly inconsistent foreign policy behaviour. 6 These three perspectives are to a degree complementary and carry some of explanatory power, but they remain incomplete. The domestic argument cannot explain why Trump changed his stance on burden-sharing relatively late in his term when the 'adults in the room' such as Defense Secretary Mattis or Chief of Staff Kellythe major constraints on Trumphad departed the administration. The structural argument fails to explain why Trump was repeatedly on the verge of withdrawing from the alliance and who the actors were that persuaded the reluctant Trump of the merits of continued support for NATO. And the psychological argument is by itself insufficient to offer a comprehensive account of Trump's NATO policy; discerning the effects of Trump's personality requires understanding how they interact with the alliance's institutional and political environment. 7 To explain Trump's puzzling NATO policy, this article incorporates but goes beyond domestic, structural, and psychological arguments by focusing on the neglected role played by NATO's Secretary General Stoltenberg and senior officials in Brussels. The omission of these actors in extant analyses is not surprising. Most scholars view NATO as a traditional military alliance, which lacks meaningful institutions and thus constitutes merely an instrument of state power. 8 Frank Schimmelfennig exemplarily observes that 'strong versions of institutional theory [which emphasize the agency of the Secretary General and wider bureaucracy] have not been prominent or supported in studies of NATO.' 9 But NATO is more than a narrow military alliance held together by common threat perceptions. NATO is a security organization, undergirded by strong institutions, interdependencies, and a shared value foundation. 10 Indeed, recent contributions affirm the growing importance of NATO senior officials, 11 which echoes an emerging wider research agenda on the significance of secretariats in fending off contestation. 12 Julia Gray, for example, shows that the quality of their bureaucracies is a key determinant of International Organization's (IO) vitality, while Maria J. Debre and Hylke Dijkstra demonstrate that IOs with greater bureaucratic capacity are less likely to die when challenged and more likely to exploit crises as opportunities for organizational growth. 13 Thus, the outcome of contestation is in many cases not predetermined but dependent on how the IO leadership responds.
Drawing on 23 original interviews with senior NATO and allied officials (both from the delegations in Brussels and capitals), this article therefore sets out to trace how the NATO Secretary General and other senior officials responded to Trump's contestation and evaluate how causally relevant these responses were for NATO's survival. It focusses on Trump's two central demands for greater transatlantic burden-sharing and closer relations with Russia. 14 The article finds that Secretary General Stoltenberg and senior NATO officials used agenda-setting and brokering strategies to overtly embrace Trump's demands for greater burden-sharing because they promised to generate most goodwill with the US President and were not harmful to the alliance. In contrast, NATO leaders used strategies of coalition-building and shielding 9 Frank Schimmelfennig, 'NATO and institutional theories of international relations', in Mark Webber and Adrian Hyde-Price, eds, Theorising NATO (London: Routledge, 2016), pp. 93-115. 10  to subtly resist Trump's calls for closer relations with Russia because they threatened to undermine NATO's raison d'être.
The empirical analysis suggests that Stoltenberg's decisive in managing the critical summit of 2018, where President Trump was on the verge of announcing a US withdrawal from NATO over burden-sharing disputes, and the Secretary General was critical in persuading Trump that allies were heeding his calls to increase their defence spending, even though increases fell short of Trump's demands. NATO leaders also helped shield NATO's defence and deterrence posture toward Russia from Trump, but here the US foreign policy establishment also played a critical role. Given the poor personal relations with Trump, no other allied leader had any noteworthy influence on the US President. Contrary to the bulk of scholarly opinion, Stoltenberg and other senior officials thus exhibited a striking degree of agency in helping NATO survive Trump.
This argument is developed as follows. First, the article theorizes how and under what conditions IOs can respond to hegemonic contestation. Second, it shows that NATO had the institutional levers, external support, and leadership to respond strategically. Third, the article traces how the NATO leadership responded to demands for greater burden-sharing and rapprochement with Russia.

Secretary generals, IO leadership, and hegemonic contestation
This section theorizes how and when IO leaders can blunt hegemonic contestation. Secretary Generals are most likely to spearhead the responses to hegemonic contestation, but they tend to be supported by other senior officials who can draw on the IO's bureaucratic machinery, including the deputy secretary general, director of the private office, or heads of divisions. In turn, hegemonic contestation is here referred to as public criticism by a hegemonic member state accompanied by demands for institutional changes and, implicit or explicit, threats of withdrawal. Hegemonic contestation poses a grave danger to the survival of IOs as they tend to be extremely dependent on, and thus vulnerable to, the hegemonic member state, which makes unrivalled material contributions and possess superior sources of influence. With its very existence in question, the IO leadership should therefore naturally seek to exploit all its formal and informal levers of power to fend off contestation. 15 For survival-seeking IO leaders, however, hegemonic contestation poses a dilemma.
On the one hand, hegemonic contestation generates enormous pressures to adapt to the hegemon's demands. In IOs where the power distribution among member is balanced, IO leaders must weigh up the potential costs of inaction in the eyes of the contesting state against the costs of adapting in the eyes of the non-contesting states, which did not pressure for change and thus appear satisfied with the status quo. 16 In IOs characterized by stark power asymmetry such as NATO, however, IO leaders must prioritize the preferences of the hegemon over those of other member states to avert potentially fatal sanctioning or withdrawal. On the other hand, hegemonic contestation often also creates heavy pressures to resist. Hegemonic contestation is likely to address core rather than peripheral features of the organization because hegemons tend to possess sufficient influence within organizations to reform technicalities. This threat to core features is reinforced when the sources of hegemonic contestation lie in the nationalist populist turn at home, which explicitly rejects the foundational multilateral principles of IOs. 17 Thus, adaptation risks undermining the very material or ideational raison d'être of the IO, which the officials tend to identify with. 18 Navigating this trade-off between potentially high costs of adaptation and resistance is a difficult task. Nonetheless, IO leaders may be able to do so if they respond strategically, which means that they proactively formulate and implement a response deliberately tailored to overtly embrace those hegemonic demands least harmful to the organization while subtly resist those deemed harmful to its integrity. Recognizing that outright resistance to the hegemon is unfeasible, IO leaders must adapt sufficiently to placate the hegemon while prioritizing certain features of the organization to protect from the hegemon's encroachment. Strategic thus contrast with passive responses of simply following orders of the hegemon, trying to sit out the contestation, or following a pre-existing playbook. 19 Whether the IO leadership can respond strategically depends on three conditions. First, secretary generals and senior officials need to have internal levers of power, here referred to as institutional capacity, otherwise they are little more than toothless administrative actors. Formulating a strategy requires sufficient size of the secretariat so that secretary generals have enough policy-grade personnel available to analyse the challenge and devise a response. 20 In order to subsequently implement the strategy, the IO leaders needs to possess formal and/or informal powers to set the agenda or take decisions. 21 The greater the level of delegation, the greater the array of potential responses available. Furthermore, IOs with public communications units should be better able to promote their strategy, particularly given the increasing mediatized environment of international politics. 22 Second, IO leaders rely on a favourable opportunity structure because they operate in a complex environment where they are rarely the most powerful actor. 23 In the case of NATO and Trump, the constellation of domestic political actors and the role of other member states is relevant. The greater the internal resistance toward the hegemonic contestation, the greater the opportunities for IO leaders to build coalitions to resist hegemonic demands. Furthermore, the role of other member states, especially the big three European NATO members Germany, France, and the United Kingdom, should not be discounted even if the IO's dependency on the hegemon is pronounced. Should other member states be in broad agreement with the NATO leadership, they will be more likely to mount a strong defence of their preferences and vice versa.
Third, aligning the latent institutional powers with external opportunities, however, is not an automatic process but requires astute leadership by secretary generals and senior officials. They need to acknowledge the existential nature of the contestation, recognize the external constraints and opportunities, and then mobilize the institutional capacity. Public administration as well as political leadership literatures highlight the significance of leader's personal qualities such as intelligence, diplomatic talent, and empathy as well as the seniority in terms of their previous positions, and thus diplomatic network and reputation among heads of states and governments, for how effectively IOs respond to contestation. 24 In addition, they 'must get along with the United States.' 25 If these three conditions are met, IO leaders can respond strategically to hegemonic contestation, which can be expressed in four mechanisms: agenda-setting, shielding, coalitionbuilding, and brokering. While these mechanisms are analytically distinct, in practice they may overlap, or actors may employ a combination of these strategies. Agenda-setting means that usually secretary generals as the most prominent officials venue-shop for the most receptive location, raise public awareness and frame issues favourable, and shape internal proceedings. 26 Shielding entails isolating the organization from the interference of the contesting hegemon by keeping a low profile, using procedural tricks to exclude controversial policies from the agenda, or precooking summit conclusions. 27 Coalition-building implies that the IO leadership cultivates close relations with like-minded actors to alter the balance of power between proponents and opponents in the IO leaders' favour. The IO leadership can variably collude with like-minded member states, orchestrate affine intermediaries to pursue their objectives, or bring non-governmental actors on board. 28 What is missing from the existing literature is that IO leaders can also build coalitions with political actors from within the contesting state if the hegemonic demands are domestically controversial. Finally, brokering consists of facilitating compromises between contesting hegemon and other member states in a way that furthers their own preferences. 29 Secretary generals frequently act as formal or informal chairs in negotiations to overcome deadlock. 30 In sum, hegemonic contestation poses an existential challenge to IOs, but IO leaders may be able to manage it if they respond strategically, which in turn depends on the personal traits of the secretary general and other leaders, the institutional machinery and powers, and the external environment.

NATO's strategic Trump management
This section analyses NATO actors' Trump management. Trump's demands for greater burden-sharing generated strong pressures for NATO leaders to adapt, while his calls for closer relations with Russia created strong pressures to resist. After briefly demonstrating that NATO met the three conditions that enable strategic responses, the following section examines how NATO actors navigated this dilemma between January 2017 and November 2020.

NATO's institutional powers, the US foreign policy establishment, and Stoltenberg's leadership
To respond strategically to Trump, NATO needs to have the institutional capacity to formulate and implement a strategic plan, find like-minded supporting actors, and benefit from astute leadership. NATO remains a largely intergovernmental organization, in which member states take decisions by unanimity in the North Atlantic Council and the International Staff and the Secretary General possess very limited decision-making authority. 31 In a formal sense, NATO's institutions are principally designed as supporting bodies for the allies. A deeper look, however, reveals that the Secretary General in particular has diplomatic and communicative 27 Derek powers at his disposal. As the permanent chair of the North Atlantic Council, he can set the agenda and facilitate compromises. He is also the organizer of NATO summits and acts as the spokesperson of the alliance. NATO ranks among the largest IOs with 1000 civilians working in the International Staff in Brussels, almost 500 of whom are policy-grade officials. 32 The International Staff includes a dedicated Public Diplomacy Division, while the Secretary General's Private Office also includes a Policy Planning Unit, an internal think tank that offers policy expertise and strategic insights.
In addition, NATO relies on like-minded actors to strategically build coalitions. Most other allies were privately in support of NATO's leaders but, as shown below, had very little influence on the US President. Political actors in the US, however, offered greater opportunities. In the US, a plethora of actors are involved in foreign policy making, including Congress, State Department, Pentagon, and the National Security Council as well as private actors including think tanks and business groups. Indeed, there was bipartisan support in Congress for the alliance, key figures in the administration like Defense Secretary Mattis were ardent champions of NATO, and so were most non-governmental actors.
Last, Jens Stoltenberg was a former prime minister of Norway before becoming NATO's Secretary General in 2014. Former heads of state tend to view themselves as equals rather than servants of what previously were colleagues in the North Atlantic Council and should have strong networks among senior politicians in member states. The recent trend toward selecting former heads of state as Secretary General is indicative of stronger diplomatic prowess of the office. 33 With a European but non-EU background, Stoltenberg was widely perceived as trusted broker without a personal agenda. 34 Moreover, Stoltenberg's deputy, Rose Gottemoeller (2016-2019) was a former US Under Secretary in the State Department with extensive connections in Washington. Thus, the three enabling conditions were sufficiently met, and NATO should thus be expected to respond strategically to Trump's contestation.

Secretary General Stoltenberg and Trump's burden-sharing demands: agenda-setting and brokering
Trump's complaints about inequitable burden-sharing dominated his discourse on NATO in the early stages of his presidency, when he went as far as to condition US collective defence guarantees on allies meeting the 2% defence spending rule. 35  up NATO.' 36 He also questioned the underlying logic of unconditional support for allies when positing that he would only defend Baltic allies against Russian aggression if they had 'fulfilled their obligations to us.' 37 The issue therefore posed a veritable threat to the very survival of NATO. Had Trump made true on his words to revoke US guarantees in case that allied defence spending did not meet his demands, this would have de facto terminated the alliance built on the principle of unconditional solidarity in face of external threats.
Trump's demands for greater transatlantic burden-sharing were largely shared by NATO actors, which had long been supportive of greater allied defence investment to meet the diverse security challenges in an increasingly hostile international landscape. 38 Adaptation would therefore not pose a threat to the integrity of NATO; the main risk for the NATO leadership lay in allies not increasing their defence spending sufficiently to satisfy Trump. As a result, they had to walk a fine line. On the one hand, they needed to publicly side with Trump and pressure allies to spend more on defence. On the other hand, they had to sell even modest increases as successes to please Trump. Indeed, senior officials were aware that allies would not immediately be able to drastically increase defence spending, given the political complexity and long-term nature of budgetary spending plans. 39 In order to simultaneously lobby allies and convince Trump, and in the absence of formal means to compel allies to increase defence spending, Stoltenberg used public communications strategies and procedural means to set the agenda and broker compromises in the background.
The Secretary General chose the public realm as his principal venue through which to pursue his strategy. In close liaison with NATO's Public Diplomacy Division, he used his prominent position to publicly pressure allies to increase defence spending and credit the US President for allegedly achieving greater burden-sharing. 40  In 2019, the Secretary General intensified his tailored communicative efforts aimed at Trump and repeatedly referred to what emerged as NATO's new mantra on burden-sharing. In the runup to Trump's state of the union speech in February, Stoltenberg appeared on Trump's favorite US news channel, Fox, crediting Trump for an 'extra $100 billion' allies will have added to their defence spending by the end of 2020. 44 When invited by Speaker of the House Nancy Pelosi as the first Secretary General of any IO to speak in front of both Houses of Congress in April 2019, he lauded President Trump's positive impact on the alliance and again referred to the burden-sharing slogan. 45 Prior to the London Leaders summit in December 2019, Stoltenberg reiterated to Trump that 'your leadership on defence spending is having a real impact', citing new defence spending figures that showed a $130 billion increase to the defence spending budgets, which is expected to rise to $400 billion by 2024. 46 Thus, the Secretary General strategically purported that Trump had prevailed over the opposition from other member states. Importantly, the Secretary General always chose to compare the spending figures to 2016the year of Trump's electionrather than 2015 when the allies' budgets first showed increases to obfuscate that reasons beyond Trump could be responsible. 47 The Secretary General not only understood the power of the media for public discourse in general and the US Presidenta reportedly avid consumer of US televisionin particular, but also consciously tailored his simplistic and servile communication style to flatter the egocentric Trump. 48 One interviewee adds that Stoltenberg would always present the defence spending figures in very simple bar charts to capture his attention and cater for Trump's alleged short attention span and inattention to detail. 49 Stoltenberg also used his procedural powers as chair of the North Atlantic Council to set the burden-sharing agenda at the most perilous moment for NATO during the Trump presidency the NATO summit in July 2018. Trump's America-First rhetoric had been particularly pronounced during that summer and in June he had refused to sign the G7 statement. Trump was also due to fly to Helsinki for a controversial bilateral meeting with President Putin right after the NATO summit and there was a distinct fear among officials that Trump could decide at short notice to skip the NATO summit. 50 While Trump attended the summit, he affirmed officials' concerns when he unleashed a personal attack on German Chancellor Merkel at a bilateral meeting on the first day of the summit (11 July). The next day, tensions escalated further, and the summit was on the verge of collapse when President Trump hijacked a working meeting originally aimed at fostering relations with Ukraine and Georgia to threaten fellow allied leaders that the US would 'go its own way' should his burden-sharing demands not be met. 51 According to one interviewee, the US delegation had 'no idea what was happening'. 52 Sensing the impending danger, Stoltenberg used his procedural power as chair of the North Atlantic Council and decided to turn the working meeting into an impromptu crisis meeting on burden-sharing. This was a highly unusual, strategic decision by the Secretary General as NATO summits tend to be ritualistic and formulaic. Calling this meeting proved critical in appeasing Trump; it played to the narcistic propensities of the US President by allowing him to vent his frustration and pressure Europeans to make concessions, before taking credit for almost all NATO reforms undertaken since 2014 in the subsequent press conference, letting him walk away with a sense of victory. 53 Complementing the agenda-setting strategy, Stoltenberg also sought to exert diplomatic pressures and broker compromises among other member states. His private office included several senior seconded officials, and he used them as ears and mouths in the capitals. 54 He also regularly toured the capitals to persuade Europeans and Canada of the need for greater defence spending. While some allies felt unease about Stoltenberg's overriding focus on defence spending as principal indicator for burden-sharing, his reputation as honest broker and skilled mediator allowed him to overcome these concerns. 55 He also tended to explicitly invoke the threat of US withdrawal to strengthen his case. 56 In November 2019, senior officials in Stoltenberg's private office helped broker a new Common Funding formula for NATO's budget. In order to alleviate Trump's criticism of allied, and in particular German, underspending, Stoltenberg's office worked behind the scenes with officials from the German chancellery and the US National Security Council to increase Germany contributions to match the reduced US contributions. 57 While this was largely symbolic given the relatively insignificant sums involved, it subsequently allowed Stoltenberg to publicly tout another victory for Trump's burden-sharing agenda.
By the end of 2019, Trump had publicly proclaimed his satisfaction at several points that 'people are paying and I'm very happy with the fact that they're paying' 58 , despite the fact that increases were well-below his demands that every ally meet the 2% rule (not to mention spending 4% of GDP on defence, a demand he made at the 2018 summit). While national defence budgets have been on the rise since 2015, only 7 out of 29 allies met the target in 2019 and only 15 had set out plans to reach 2% of GDP of overall defence spending by 2024. 59 There is therefore a correlation between Trump's conversion on burden-sharing and Stoltenberg's' strategic responses; indeed, there are several pieces of evidence that suggest that the Secretary General and senior officials played a causal role.
First, Trump's own comments suggest that Stoltenberg was critical in persuading him. Throughout his tenure, NATO's Secretary General maintained an amicable relationship with Trump, a rarity for any leader. 60 Trump heaped lavish praise on him, describing his relationship with Stoltenberg as 'outstanding' and stating that he had 'done an excellent job', which was reinforced when Trump supported extending Stoltenberg's term as Secretary General for another two years. Crucially, he established a direct link between Stoltenberg and burdensharing, exclaiming that 'the media never gives me credit but he gave me credit, now we're up to way over $100 billion.' 61 Moreover, Stoltenberg's $100 billion slogan evidently gained traction with Trump, especially after the Secretary General trumpeted the slogan on Fox News, which Trump immediately retweeted and then cited for the first time in his state of the union address only one week later. 62 Second, several closely involved officials confirm that Trump began changing his stance after encounters with Stoltenberg, whose adroit flattery pushed the right buttons with Trump. 63 Close observers confirm this impression, with the then UK ambassador in Washington characterizing Stoltenberg as the 'master Trump-whisperer' 64 , while one interviewee noted that the Secretary General was always one of the first points of contact when defence decisions were impending and that 'Trump looked to Stoltenberg for advice '. 65 Third, the beginning of Trump's conversion can be traced to the 2018 summit, which was a critical juncture for the alliance. The outcome of the summit was highly contingent; US officials at the summit feared that Trump would announce the US withdrawal from NATO at the press conference on 12 July and had even instructed lawyers to analyse NATO's founding Treaty for advice on the legal mechanisms (though Congress would have prevented a formal withdrawal). Secretary Mattis was strikingly absent from the stage when Trump gave the press conference and expressed his willingness to resign that day. 66 Without Stoltenberg's unscripted and spontaneous decision to call the emergency session, all indicators suggests that Trump would at least have caused severe damage to the alliance.
And fourth, the principal alternative explanation that the US foreign policy establishment, or other allies, tamed Trump cannot account for the conversion, which began with the 2018 summit and culminated in his first public embrace of NATO's turnaround on defence spending in February 2019. By then, however, the 'adults in the room' had long lost influence or even left the administration. 67 68 One directly involved official confirmed that they 'had no intellectual impact on Trump' and 'never made a dent' in Trump's views on NATO. 69 Other NATO allies helped passively persuade Trump by moderately increasing their defence budgets, which NATO actors could then exploit, but they too had little direct influence on the President. 70 While then UK Prime Minister May successfully extracted a vague commitment to NATO from Trump at their first bilateral meeting in January 2017, their relationship quickly soured when Trump openly criticised her approach to Brexit. 71 Similarly, the good relations between Trump and French President Macron were short-lived. With German chancellor Merkel, Trump appeared to have a personal feud, although she helped Stoltenberg manage Trump at the 2018 summit by stoically enduring his attacks. 72 One official stressed that Stoltenberg was 'the only one in Europe who had Trump's ear'. 73 In sum, this section demonstrates ample evidence that Stoltenberg not only responded strategically to Trump's contestation by tailoring their public agenda-setting to the idiosyncrasies of the US president and adroitly employing procedural powers to manage the 2018 summit. It also suggests that these activities were a causal factor in eventually persuading Trump that NATO were heeding his calls for greater transatlantic burden-sharing, despite very limited affirmative evidence. Stoltenberg's personal leadership was critical, which was manifest in his conscious decision to build a close personal rapport with Trump and his diplomatic skill to do so. He also understood the power of the media for shaping Trump's thinking and used his procedural powers as chair most effectively during the 2018 summit to avert the worst-case scenario. Several interviewees observed that Stoltenberg's predecessor, Anders Fogh Rassmusen, would have never managed to handle Trump given his allegedly more pronounced ego. 74 Trump's narcistic disposition and vulnerability to flattery was also a necessary for NATO's agenda-setting strategy to succeed.
In fact, Stoltenberg and the NATO leadership employed similar strategies in the case of China. Senior officials were aware how much importance the Trump administration attached to putting China on NATO's agenda but also of the reluctance of many European allies to militarise relations and distract NATO's focus from Russia. 75 Placating Trump required walking the fine line again of setting the agenda on China while subsequently selling the mere reference to China in the 2019 London Declaration as substantial progress. After Stoltenberg helped broker a compromise at the summit, 76 allies 'recognise[d] that China's growing influence and international policies present both opportunities and challenges'. 77 It was thus another incident of NATO demonstrating ostensible responsiveness to the US President even if the declaration implied no operational consequences. 78

Stoltenberg and NATO's Russia policy: coalition-building and shielding
During the presidential election and when in office, Trump made a string of interventions that suggested he wanted to re-establish cordial relations with Russia. He repeatedly expressed his admiration for Putin, calling him a 'strong leader, a powerful leader' and that 'I would love to be able to get along with Russia'. 79 He also lobbied to reintegrate Russia in the G7, denied Russian interference in the 2016 US election, implicitly acknowledged Russia's annexation of Crimea, and cast doubts about whether his administration would uphold the sanction regime. 80 Moreover, his foreign policy team and trusted circle was stacked with Russophiles with close connections to Moscow (including Paul Mannafort, Carter Page, and Newt Gingrich). 81 Since Russia's annexation of Crimea and invasion of Eastern Ukraine in 2014, however, NATO had shifted back to its original purpose of providing territorial security to its member states against Russia. Hence, Trump's ambiguous benevolence set off alarm bells and NATO and national officials worried that Trump would reduce the US investment in Europe and subvert NATO initiatives. 82 One official said that the 'Russia question was looming large over Brussels'. 83 Such a dilution, or even reversal of NATO's defence and deterrence posture toward Russia would have undermined the very raison d'être of the alliance originally founded to defend European against Soviet aggression. NATO leaders were steadfast in their opposition toward thawing relations with Russia but were caught in the dilemma of hegemonic contestation that rendered overt resistance to Trump unfeasible. Instead, the NATO leadership set out to subtly resist Trump's demands by building coalitions with favourable actors in the US and shielding Russia policy from Trump.
The case of Russia policy involved a different logic than the case of burden-sharing. Rather than being a matter of political deliberation, NATO's stance on Russia was a matter of concrete policy making, which necessarily lend the US foreign policy bureaucracy a much greater role. Stoltenberg sought to exploit the political support NATO enjoyed among the US foreign policy establishment by building coalitions with supportive actors to coordinate policy and maintain US domestic support for the alliance. 84 Relying on his personal network as much as on his deputy's, the Secretary General worked through two channels in with Trump. 88 Throughout Trump's term in office, the Secretary General also cultivated relations with US parliamentarians. He regularly hosted US delegations in Brussels and spoke as first Secretary General of any IO in front of both Houses of Congress in April 2019 to rally domestic support for NATO.
Besides coalition-building, the NATO leadership also sought to shield NATO's Russia policy from Trump. In an attempt to keep the issue beneath Trump's radar, Stoltenberg prioritized burden-sharing over Russia policy in his public communications with Trump. 89 In the press conferences or remarks following their six bilateral meetings between April 2017 and December 2019, Stoltenberg always emphasized the need for greater burden-sharing. However, he did not mention Russia policy in three of the press conferences, while in the others he only addressed Russia cursorily. 90 By selectively engaging with Trump's demands in public, the Secretary General tried to focus his attention on the agenda NATO leaders supported and were least harmful to the organization, which illustrates that NATO's response to Trump's demands on burden-sharing and Russia were intimately connected.
The most visible and consequential instance of shielding occurred in the run-up to the summit in July 2018. In light of Trump's refusal to sign the G7 communique, senior officials tried to Trump-proof the summit. The Secretary General together with US diplomats successfully pressured ambassadors to agree upon a declaration prior to the summit to avoid last-minute interferences from Trump. 91 They decided to keep the text 'short and sweet' and publicly downplay the achievements to keep them beneath Trump's radar. 92 After Stoltenberg successfully appeased Trump, allies agreed on the Readiness Initiative, criticized Russia, invited North Macedonia to join, and established an Atlantic Command post to facilitate swift response to a potential war in Europe. Further manifestations of NATO actor's shielding efforts were their decisions to re-organize summits where Russia policy was a central discussion point. One of the first decisions of the Secretary General was to postpone Trump's first visit to NATO's headquarters to May 2017. Officials in the International Staff hoped that Trump's anger would dissipate over the months and that he would be taught the value of NATO by the 'adults in the room', i.e. General Mattis and his then chief of staff John Kelly. 93 In an attempt to prevent Trump disrupting celebrations, the Secretary General also downgraded NATO's 70 th anniversary summit in April 2019 in Washington D.C. to a foreign ministerial meeting, which was attended by Secretary of State Pompeo instead. 94 NATO maintained a robust defence and deterrence posture toward Russia throughout Trump's tenure, despite the US President's calls to the contrary. 95 NATO's new Readiness Initiative would, if implemented, be a significant step toward preparedness for conflicts with Russia. The US also steadily increased the budgetary allocation for the European Deterrence Initiative, 96 actively participated in military exercises (20 000 US troops participated in the Trident Juncture exercise in 2018) and took command of one multinational battlegroup in Poland under the umbrella of the Enhanced Forward Presence. Notwithstanding Trump's announcement to withdraw almost 12 000 troops from Germanyplans which have since been put on ice by the Biden Administration -Trump did not substantially undermine NATO's posture vis-à-vis Russia. 97 Where the NATO leadership played a causal role in converting Trump on burden-sharing, they played a lesser if still meaningful part in maintaining a robust Russia policy. The effects of Stoltenberg's strategic neglect of Russia in his public communications are difficult to discern empirically but it probably helped distract Trump. The most consequential episode of agency of NATO actors was their shepherding of the 2018 summit, which proved to be among the most consequential in years in terms of new policy measures on Russia. Nevertheless, the unwavering commitment to NATO in the Pentagon, the State Department, the National Security Council, and Congress was essential in resisting Trump's demands for closer relations with Russia.

Conclusion
NATO survived one of the gravest contestations in its history, which was not a predetermined outcome. This article departed from the puzzling observations that President Trump, first, changed his stance on transatlantic burden-sharing despite increases in allied defence budgets remaining significantly below his demands and, second, that the US reinforced NATO's defence and deterrence posture toward Russia notwithstanding Trump's preferences to the contrary. Inspired by flaws in existing accounts of Trump's NATO policy and an emerging research agenda on the agency of IO leadership under pressure, this article examined the role played by the NATO leadership, in particular Secretary General Stoltenberg. It traced how NATO actors responded to Trump by setting the burden-sharing agenda, shielding Russia policy from the US President, building coalitions with favourable actors in the US foreign policy machinery, and brokering compromises among allies in the background. The empirical section demonstrated that NATO leaders strategically navigated the dilemma of hegemonic contestation by signalling sufficient adaptation to placate the US President while protecting core features of the alliance from Trump's demands. 95 Interview 7; Interview 12; and interview 14. 96 The funding increased from $800 million in 2016, to $3.4 billion in 2017, to $4.8 billion in 2018, to $6.5 billion in 2019, to $5.9 billion in 2020. 97 Jens Ringsmose and Mark Webber, "Hedging their bets? The case for a European pillar in NATO," Defence Studies 20: 4, 2020, pp. 295-317.
Whether Trump truly changed his mind on NATO is impossible to ascertain, but Stoltenberg and senior officials were a necessary factor in convincing Trump to change position on burdensharing and managing the critical 2018 summit, while they played a supportive role in maintaining a robust Russia policy. Thus, NATO leaders exhibited a striking degree of agency in helping NATO survive. Indeed, had NATO leaders acted as passive servants, as most writings presume, Trump could have been much more destructive. Congress would have prevented Trump from formally withdrawing the US from NATOin 2019, both the House of Representatives and the Senate passed legislation to prevent Trump from using federal funds to withdraw from NATO. 98 But without the actions of Stoltenberg, it is plausible that Trump had still expressed his intentions to withdraw, the mere announcement of which would have de facto meant the end of NATO in its current form. A passive secretary general would have been unable to build a close rapport with Trump and convince him in 2019 that he had successfully imposed his will onto allies to achieve greater burden-sharing.
In addition, servile NATO actors would have been unable to manage the highly contingent 2018 summit. National officials would have still tried to shield the proceedings from Trump. But without senior officials heaping pressures on allies to agree to a summit declaration early and, crucially, without the Secretary General strategically using his procedural powers to provide Trump the floor to voice his grievances, the US President would have likely refused to sign the summit declaration like he did at the G7 and in doing so blocked important NATO initiatives and massively undermined the credibility of the alliance. It is entirely conceivable that he would have continued publicly berating allies, toying with the idea of withdrawing, and potentially politicize NATO membership during the 2020 presidential election campaign.
The analysis broadly vindicates the theoretical claim that IO leaders can protect their preferences, even amid hegemonic contestation. It demonstrated that this strategic capacity was contingent on Stoltenberg's astute leadership, IO leaders' often overlooked institutional levers of power such as the agenda-setting powers of the chair of the North Atlantic Council, and strong support for the alliance from the US foreign policy establishment. Other allies played a minor role in directly managing Trump, albeit without their moderate increases in defence spending, even the 'master Trump-whisperer' Stoltenberg would have been unable to influence Trump. The fact that these defence spending increases had likely not happened without Russian aggressions in Ukraine constitutes one of the many fortunate contingencies for NATO. 99 On a broader level, the article's counterintuitive findings on NATO's agential qualities should nuance the prevalent state-centric view of NATO. Some insights are probably specific to the Presidency of Trump; his particular susceptibility to flattery and inattention to policy details, which helped NATO's strategic behaviour, appear seldom if not unique among heads of state or government. However, NATO has gradually become a stronger institution and the office of the Secretary General has also been empowered. The case of Trump's contestation testifies that NATO is a resilient security organization with greater power than is traditionally attributed to it. This should be particularly relevant because the fundamental debate about the future of the alliance shows no sign of abating. French President Macron's allegation that NATO was 'braindead' highlights persisting tensions in the alliance. Moreover, the rise of China, enduring conflicts among some of the allies, democratic decay across the West, or new security challenges to name but a few will raise question about the relevance of NATO, but the NATO leadership has demonstrated that they will not be passive pawns in the game.

Funding information
This article is part of a project that has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (grant agreement No 802568).