Why NATO survived Trump: the neglected role of Secretary-General Stoltenberg

International Affairs 97: 6 (2021) 1863–1881; doi: 10.1093/ia/iiab167 © The Author(s) 2021. Published by Oxford University Press on behalf of The Royal Institute of International Affairs. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/ licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. NATO only just survived the presidency of Donald Trump. Once in office, Trump—who had distinguished himself from virtually all US presidents since the Second World War in his active hostility towards the alliance during the presidential campaign—repeatedly toyed with the idea of withdrawing from NATO, and was on the verge of doing so publicly at the 2018 NATO summit.1 But while the president withdrew the United States from the Iran Nuclear Deal, the Paris Climate Agreement and UNESCO, and undermined the WTO, WHO, UN Refugee Agency and Green Climate Fund from within, he 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’.2 Given that the United States is the de facto indispensable power within the alliance, the intuitive explanation for NATO’s survival would be that it successfully adapted to Trump’s demands. However, the empirical record suggests that NATO only partially adapted to Trump’s demands for greater transatlantic burden-sharing and resisted his calls for closer relations with Russia. Two specific questions therefore emerge. First, why did Trump change his stance on transatlantic burden-sharing, even though increases in allied defence spending remained

International Affairs 97: 6, 2021 significantly below his demands? Second, why did the United States go so far as to reinforce NATO's defence and deterrence posture vis-à-vis Russia, despite Trump's calls to the contrary? 3 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 locates 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 within liberalism, allegedly entrenched in US society, that prevent a constrained foreign policy and the withdrawal of support for NATO. 4 The second camp emphasizes that continued support for NATO is the rational utility-maximizing 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 have some 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 Kelly-the major constraints on the president-had departed the administration. The structural argument fails to explain why Trump was repeatedly on the verge of withdrawing from the alliance and who were the actors persuading 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 the domestic, structural and psychological arguments to focus on the neglected role played by NATO's Secretary-General, Jens Stoltenberg, and senior officials in Brussels. The omission of these actors from analyses to date is not surprising. Most scholars view NATO as a traditional military alliance, 3  which lacks meaningful institutions and thus constitutes merely an instrument of state power. 8 As Frank Schimmelfennig observes, 'strong versions of institutional theory [which emphasize the agency of the Secretary-General and the 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: it is a security organization, undergirded by strong institutions, interdependencies and shared foundational values. 10 Indeed, recent contributions to the scholarship affirm the growing importance of NATO senior officials, 11 echoing an emerging wider research agenda on the significance of secretariats in fending off contestation. 12 Julia Gray, for example, shows that the quality of its bureaucracy is a key determinant of the vitality of an international organization (IO), while Maria 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 (from delegations in Brussels and national capitals), this article sets out to trace how the NATO Secretary-General and other senior officials responded to Trump's contestation and to evaluate how causally relevant these responses were for NATO's survival. 14 It focuses on Trump's two central demands: for greater transatlantic burden-sharing and for closer relations with Russia. 15 General Stoltenberg and senior NATO officials used overt agenda-setting  and brokering strategies to 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 subtle strategies of  coalition-building and shielding to 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 played a decisive role 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 that his contribution was also critical in persuading Trump that allies were heeding his calls to increase their defence spending, even though those increases fell short of Trump's demands. NATO leaders helped shield NATO's defence and deterrence posture towards Russia from Trump, but here the US foreign policy establishment played a critical role too. Given the poor personal relations between allied leaders and Trump, none of them 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.
The 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, it 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, such as the deputy secretary-general, director of the private office or heads of divisions. Hegemonic contestation here refers to 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 possesses superior sources of influence. When the IO's very existence is called into question, its leadership should therefore naturally seek to exploit all its formal and informal levers of power to fend off contestation. 16 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 members is balanced, IO leaders must weigh up the potential costs of inaction in the eyes of International Affairs 97: 6, 2021 the contesting state against the costs of adapting in the eyes of the non-contesting states, which have not exerted pressure for change and thus appear satisfied with the status quo. 17 However, in IOs characterized by stark power asymmetry-such as NATO-leaders must prioritize the preferences of the hegemon over those of other member states to avert potentially fatal sanctions 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. 18 Thus, adaptation risks undermining the very material or ideational raison d'être of the IO, with which officials tend to identify. 19 Navigating this trade-off between the potentially high costs of both adaptation and resistance is a difficult task. Nonetheless, IO leaders may be able to do so if they respond strategically; that is, if they proactively formulate and implement a response deliberately tailored to overtly embrace those hegemonic demands least harmful to the organization while subtly resisting those deemed harmful to its integrity. Recognizing that outright resistance to the hegemon is infeasible, IO leaders must adapt sufficiently to placate the hegemon while prioritizing certain features of the organization to protect them from the hegemon's encroachment. Strategic responses thus contrast with passive responses, which may take the form of simply following the orders of the hegemon, trying to sit out the contestation or following a pre-existing script. 20 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 a secretariat of sufficient size to provide secretary-generals with enough policy-grade personnel (i.e. not linguists, administrative assistants or manual workers) available to analyse the challenge and devise a response. 21 In order to go on to implement the strategy, the IO's leaders need to possess formal and/or informal powers to set the agenda or take decisions. 22  array of potential responses available. Furthermore, IOs with public communications units should be better able to promote their strategy, particularly given the increasingly mediatized environment of international politics. 23 Second, IO leaders rely on a favourable opportunity structure because they operate in a complex environment where they are rarely the most powerful actors. 24 In the case of NATO and Trump, the constellation of US domestic political actors and the role of other member states are relevant. The greater the internal resistance towards the hegemonic contestation, the greater the opportunities for IO leaders to build coalitions to resist hegemonic demands. Furthermore, the role of other member states-in the case of NATO, especially the big three European 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 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 does not happen automatically 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. Both the public administration and political leadership literatures highlight the significance of leaders' personal qualities, such as intelligence, diplomatic talent and empathy, as well as their seniority in terms of their previous positions, and thus their diplomatic networks and reputation among heads of states and governments, in influencing how effectively IOs respond to contestation. 25 In addition, they 'must get along with the United States'. 26 If these three conditions are met, IO leaders can respond strategically to hegemonic contestation. Such a response can be expressed in four mechanisms: agendasetting, shielding, coalition-building and brokering. While these mechanisms are analytically distinct, in practice they may overlap, or may be employed in combination. Agenda-setting usually involves secretary-generals, as the most prominent officials, venue-shopping for the most receptive location, raising public awareness and framing issues favourably, and shaping internal proceedings. 27 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 preparing summit conclusions in advance to lay out the tracks. 28 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. In pursuit of this end, the IO leadership can variously collude with like-minded member states, orchestrate closely connected intermediaries in pursuit of shared objectives, or bring non-governmental actors on board. 29 What is missing from the existing literature is recognition 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 the contesting hegemon and other member states in a way that furthers the secretariat's own preferences. 30 Secretary-generals frequently act as formal or informal chairs in negotiations to overcome deadlock. 31 In sum, hegemonic contestation poses an existential challenge to IOs, but IO leaders may be able to manage it if they respond strategically. Success in this enterprise depends in turn on the personal traits of the secretary-general and other leaders, the institutional machinery and powers of the particular IO, and the external environment.

NATO's strategic management of Trump
This section analyses NATO actors' management of President Trump. 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 needed 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. 32 In a formal sense, NATO's institutions are 28 Derek Beach and Sandrino Smeets, 'The unseen hands principally designed as supporting bodies for the allies. A closer look, however, reveals that the secretary-general in particular has diplomatic and communicative 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 1,000 civilians working in the international staff in Brussels, almost 500 of whom are policy-grade officials. 33 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 build strategic 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 United States, however, offered greater opportunities for coalition-building. In the US, a plethora of actors are involved in foreign policy-making, including Congress, the State Department, the 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 such as Defense Secretary Mattis were ardent champions of NATO, and so were most non-governmental actors.
Finally, 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 those who were previously colleagues in the North Atlantic Council, and generally have strong networks among senior politicians in member states. The recent trend towards selecting a former head of state as secretary-general is indicative of the increasing diplomatic status of the office. 34 With a European but non-EU background, Stoltenberg was widely perceived as trusted broker without a personal agenda. 35 Moreover, Stoltenberg's deputy from 2016 to 2019, Rose Gottemoeller, 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 could consequently 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 so far as to make US collective defence guarantees conditional on allies meeting the 2 per cent defence 33  Trump's demands for greater transatlantic burden-sharing were largely shared by NATO's institutional actors, which had long been supportive of greater allied defence investment to meet the diverse security challenges in an increasingly hostile international landscape. 39 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 side with Trump in public and put pressure on 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 increase defence spending sharply, given the political complexity and long-term nature of budgetary allocations and spending plans. 40 In order to lobby allies and simultaneously 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 the principal venue through which to pursue his strategy. In close liaison with NATO's public diplomacy division, he used his prominent position to put public pressure on allies to increase their defence spending and credit the US president for allegedly achieving greater burden-sharing. 41  spending, pledging to 'work with President Trump on how to adapt NATO'. 42 On his first visit to Washington in April 2017, Stoltenberg embraced Trump's criticism of allies' insufficient defence spending, and also expressed gratitude to Trump for his 'strong commitment to Europe'. 43 And appeasing Trump and playing to his ego seemed to be the purpose of Stoltenberg's visit to the White House in May 2018, when he thanked the US President for his 'leadership ... on the issue of defence spending [which] has really helped to make a difference', a sentiment he echoed at the Brussels summit in July 2018. 44 In 2019, the Secretary-General intensified his tailored communicative efforts aimed at Trump, repeatedly referring to what emerged as NATO's new mantra on burden-sharing. In the run-up to Trump's State of the Union speech in February, Stoltenberg appeared on the President's favourite US news channel, Fox, crediting him for an 'extra $100 billion' allies would have added to their defence spending by the end of 2020. 45 When invited by Speaker of the House Nancy Pelosi to be 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. 46 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 members' defence spending budgets, expected to rise to $400 billion by 2024. 47 Thus, the Secretary-General strategically promoted the view that Trump had prevailed over opposition from other member states. Importantly, the Secretary-General always chose to compare the spending figures to 2016the year of Trump's election-rather than 2015, when the allies' budgets first showed increases, to obscure the possibility that factors other than Trump could be responsible. 48 The Secretary-General not only understood the power of the media in public discourse in general and for the US president-reportedly an avid consumer of US television-in particular, but also consciously adopted a simplistic and servile communication style to flatter the egocentric Trump. 49 One International Affairs 97: 6, 2021 interviewee adds that Stoltenberg would always present the defence spending figures in very simple bar charts to capture the president's attention and cater for his alleged short attention span and inattention to detail. 50 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 Vladimir Putin right after the NATO summit, and there was a distinct fear among officials that he might decide at short notice to skip the NATO summit. 51 While in the event Trump did attend the summit, he bore out officials' concerns when he unleashed a personal attack on German Chancellor Angela 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 United States would 'go its own way' should his burdensharing demands not be met. 52 According to one interviewee, the US delegation had 'no idea what was happening'. 53 Sensing the impending danger, Stoltenberg used his procedural powers as chair of the North Atlantic Council 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. It proved to be of critical importance in appeasing Trump, playing to the narcissistic propensities of the US president by allowing him to vent his frustration and put pressure on 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. 54 Alongside the agenda-setting strategy, Stoltenberg also sought to exert complementary 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. 55 He also regularly toured those 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 the principal indicator of burden-sharing, his reputation as an honest broker and skilled mediator allowed him to overcome these concerns. 56  tended explicitly to invoke the threat of US withdrawal to strengthen his case. 57 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, under-spending, Stoltenberg's office worked behind the scenes with officials from the German chancellery and the US National Security Council to increase German contributions to match the reduced US level. 58 While this was largely symbolic, given the relatively insignificant sums involved, it subsequently allowed Stoltenberg to claim another public victory for Trump's burden-sharing agenda.
By the end of 2019, Trump had proclaimed his satisfaction at several points that 'people are paying and I'm very happy with the fact that they're paying', 59 despite the fact that increases were well below his demand that every ally meet the 2 per cent target (not to mention spending 4 per cent of GDP on defence, a demand he made at the 2018 summit). While national defence budgets have been on the rise since 2015, only seven out of 29 allies met the target in 2019 and only 15 had set out plans to reach 2 per cent of GDP of overall defence spending by 2024. 60 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 in that conversion.
First, Trump's own comments suggest that Stoltenberg played a critical role in persuading him. Throughout his tenure, NATO's Secretary-General maintained an amicable relationship with Trump, a rarity for any leader. 61 Trump heaped lavish praise on Stoltenberg, describing their relationship as 'outstanding' and stating that he had 'done an excellent job'-a judgement 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'. 62 Moreover, Stoltenberg's $100 billion slogan evidently gained traction with Trump, especially after the Secretary-General trumpeted it on Fox News, an announcement which Trump immediately retweeted and then cited for the first time in his State of the Union address just one week later. 63 Second, several closely involved national and NATO officials confirm that Trump began changing his stance after encounters with Stoltenberg, whose adroit 57 Interview 2; interview 13; interview 18. flattery pushed the right buttons with the president. 64 Close observers confirm this impression: the then British Ambassador in Washington characterized Stoltenberg as the 'master Trump-whisperer', 65 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'. 66 Third, the beginning of Trump's conversion can be traced to the 2018 summit, which was a critical moment for the alliance. The outcome of the summit hung in the balance; US officials attending 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, in private, implied his willingness to resign that day. 67 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 his 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, the 'adults in the room' had long lost influence in the administration, and some had even left. 68 Secretary of State Tillerson was fired in March 2018, and National Security Advisor McMaster resigned in April 2018. Defense Secretary Mattis and Chief of Staff Kelly resigned in December 2018 and January 2019 respectively, and had reportedly lost the president's ear long before their departures. 69 One official directly involved with the US president confirmed that they 'had no intellectual impact on Trump' and 'never made a dent' in his views on NATO. 70 Other NATO allies helped to bring Trump round in a more passive way by moderately increasing their defence budgets, a point which NATO actors could then exploit, but they too had little direct influence on the President. 71 While the then British prime minister Theresa 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 criticized her approach to Brexit. 72 Similarly, good relations between Trump and French President Macron were short-lived. With International Affairs 97: 6, 2021 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. 73 One official stressed that Stoltenberg was 'the only one in Europe who had Trump's ear'. 74 In sum, there is ample evidence that Stoltenberg responded strategically to Trump's contestation by tailoring NATO's public agenda-setting to the idiosyncrasies of the US president and adroitly employing procedural powers to manage the 2018 summit. Indeed, it appears that these activities were a causal factor in eventually persuading Trump that NATO was heeding his calls for greater transatlantic burden-sharing, despite very limited affirmative evidence. A critical role here was played by Stoltenberg's personal leadership, as manifest in his conscious decision to build a close personal rapport with Trump and his diplomatic skill in doing so. He also understood the power of the media in 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 Rasmussen, would have never managed to handle Trump, given his allegedly more pronounced ego. 75 Trump's narcissistic disposition and vulnerability to flattery were also key factors in the success of NATO's agenda-setting strategy.
Interestingly, 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 militarize relations with Beijing and distract NATO's focus from Russia. 76 Placating Trump required once again walking a fine line 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, 77 allies 'recognise[d] that China's growing influence and international policies present both opportunities and challenges'. 78 This represented another incidence of NATO demonstrating ostensible responsiveness to the US President, even if the declaration implied no operational consequences. 79

Stoltenberg and NATO's Russia policy: coalition-building and shielding
Both during the presidential campaign 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 saying 'I would love to be able to get along with 73  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, prompting concern among NATO and national officials that Trump would reduce the US investment in Europe and subvert NATO initiatives. 83 One official said that the 'Russia question was looming large over Brussels'. 84 Such a dilution, or even reversal, of NATO's defence and deterrence posture towards Russia would have undermined the very raison d'être of the alliance originally founded to defend Europeans against Soviet aggression. NATO leaders were steadfast in their opposition to a thaw in relations with Russia, but were caught in the dilemma of hegemonic contestation that rendered overt resistance to Trump unfeasible. Consequently, the NATO leadership set out to resist Trump's demands in subtle fashion by building coalitions with favourable actors in the United States and shielding Russia policy from Trump.
The case of Russia policy involved a different logic from the case of burdensharing. Rather than being a matter of political deliberation, NATO's stance on Russia was a matter of concrete policy-making, which necessarily gave 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. 85 Relying on both his own and his deputy's personal network, the Secretary-General worked through two channels in the US system: first, the traditional transatlantic establishment in the State Department, Pentagon and parts of the National Security Council; and second, Congress. Trump's initial appointments of NATO-sceptical officials and foreign policy advisers proved short-lived, and transatlanticist establishment figures soon took over. criticized Russia, invited North Macedonia to join, and established an Atlantic Command post to facilitate swift response to a potential war in Europe. NATO actors' efforts at shielding were further manifest in their decisions to reorganize summits where Russia policy was a central discussion point. One of the Secretary-General's first decisions was to postpone Trump's first visit to NATO's headquarters to May 2017. Officials in the alliance's 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', that is, General Mattis and his then chief of staff John Kelly. 94 In an attempt to prevent Trump from disrupting the celebrations, the Secretary-General also downgraded NATO's 70th anniversary summit in April 2019 in Washington DC to a foreign ministerial meeting, which was attended by Secretary of State Mike Pompeo instead. 95 NATO maintained a robust defence and deterrence posture towards Russia throughout Trump's tenure, despite the US President's calls to the contrary. 96 NATO's new Readiness Initiative would, if implemented, be a significant step towards preparedness for conflict with Russia. The United States also steadily increased the budgetary allocation for the European Deterrence Initiative, 97 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 of the withdrawal of almost 12,000 troops from Germany-plans which have since been put on ice by the Biden administration-Trump did not substantially undermine NATO's posture vis-à-vis Russia. 98 Whereas NATO leaders played a causal role in converting Trump on burdensharing, they played a lesser but still significant 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 to distract Trump. The most consequential episode of agency on the part 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. Whether Trump truly changed his mind on NATO is impossible to ascertain, but Stoltenberg and senior officials were a necessary factor in persuading him to change his public position on burden-sharing and in managing the critical 2018 summit, while also playing 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 writing on the subject presumes, Trump could have been much more destructive. Congress would have prevented Trump from formally withdrawing the United States from NATO-in 2019, both the House of Representatives and the Senate passed legislation to prevent him from using federal funds to bring about such a withdrawal. 99 But without the actions of Stoltenberg, it is plausible that if Trump had still expressed his intention to withdraw, the mere announcement of that intention would have meant the de facto end of NATO in its current form. A passive secretary-general would have also been unable to build a close rapport with Trump and convince him in 2019 that he had successfully imposed his will on allies to achieve greater burden-sharing.

Conclusion
In addition, passive 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 with the opportunity to voice his grievances, the US President would probably have refused to sign the summit declaration (as 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 withdrawal, and potentially politicizing NATO membership during the 2020 presidential election campaign.
International Affairs 97: 6, 2021 The analysis presented here broadly vindicates the theoretical claim that IO leaders can protect their preferences, even under conditions of hegemonic contestation. It has demonstrated that this strategic capacity was contingent on Stoltenberg's astute leadership, the often overlooked institutional levers of power at the disposal of IO leaders (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, though without their moderate increases in defence spending, even the 'master Trump-whisperer' Stoltenberg would have been unable to influence the president. The fact that these defence spending increases would probably not have happened in the absence of Russian aggression in Ukraine constitutes one of the many fortunate contingencies for NATO. 100 On a broader level, the article's counter-intuitive findings on NATO's agential qualities should moderate the prevalent state-centric view of NATO. Some insights are probably specific to the Trump presidency; his particular susceptibility to flattery and inattention to policy detail, which helped NATO's strategic behaviour, appear seldom if ever among heads of state or government. However, NATO has gradually become a stronger institution, and the office of the secretarygeneral 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 observation is particularly relevant because the fundamental debate about the future of the alliance shows no sign of abating. French President Macron's allegation that NATO is 'brain-dead' highlights persisting tensions in the alliance. 101 However, while the rise of China, enduring conflicts among some of the allies, democratic decay across the West and new security challenges, to name but a few factors, will raise questions about the relevance of NATO, its leaders have demonstrated that they will not be passive pawns in the game.