Whose balance? A constructivist approach to balance of power politics

Abstract Realist and rationalist approaches to balance of power politics underplay the degree to which balances are socially constructed. We develop a constructivist approach that accounts for the elusive and contentious nature of the balances that states seek to balance. The approach foregrounds contests over balance interpretations between states that shape whose conceptions and assessments underpin the making and remaking of balances. We argue that shifts in balancing practices are crucial to the dynamics of these contests. To substantiate this argument, we empirically study the contests over the conventional and theatre nuclear balances between East and West in the last two decades of the Cold War. The case study shows that the conventional and theatre nuclear arms control negotiations – that is, the partial shift from adversarial to associational balancing – initially fuelled and amplified both contests. At the same time, the arms control negotiations eventually resolved the contests through the development of shared understandings of the two balances, thus ending the adversarial balancing between East and West. Overall, we contribute to theory development on balance of power politics by highlighting the importance of contests over balance interpretations and by providing insights into the politics and dynamics that shape these contests.


Introduction
The balance of power has been a prominent concept and practice in world politics for several centuries. Over time, both the concept and the practice have been associated with multiple meanings. 1 In a broad understanding, speaking of a balance of power means to highlight the characteristics of the distribution of power among a set of actors. In a more narrow sense, the balance of power denotes a particular form of this distribution: a power equilibrium among the most powerful (groups of) states that ensures that none of them can dominate the others.
In this article, we argue that balance of power politics cannot be fully understood without studying the contests over the interpretation of the balance of power that shape which balance it is that the states involved seek to balance. Research on balance of power politics is dominated by realist approaches. 2 These approaches generally treat the distribution of power as a distribution and resolution of such contests. Although power is a multidimensional concept, 10 the literature on balance of power politics often focuses on one particular dimension, namely military capabilities and power. 11 We consequently selected a case study that allows to demonstrate that even 'hard' distributions of power such as military balances are subject to contests over their conception and interpretation: the Cold War disputes over the conventional and theatre nuclear balances in the 1970s and 1980s. We show that the partial shift from adversarial to associational balancing gave rise first to the dispute over the conventional balance and then the dispute over the theatre nuclear balance. The arms control negotiations primed the themes of the two conteststhat is, the balance conceptions and interpretations that were under disputeand contributed to their eventual resolution in the second half of the 1980s. We then broaden the research agenda beyond contests among rivalling states. In the third part, we discuss the two levels of balance of power politicsthe development of balancing strategies within states (or alliances) and the interplay of the balancing strategies of competing states (or alliances)and illustrate that contests over balance interpretations may also occur within alliances. In the fourth part, we highlight that the two levels take place in a transnational arena in which states compete with non-state actors over the interpretation of balances.

Comparative practices in balance of power politics
In this section, we develop a constructivist framework for studying the comparative practices that shape balance of power politics. We conceptualise balance of power politics as a repertoire consisting of comparative, adversarial, and associational practices (first subsection), argue for a stronger analytical focus on contests over balance conceptions and interpretations (second subsection), and discuss why associational balancing tends to initially intensify the contests (third subsection).

A repertoire of practices
As Emanuel Adler and Patricia Greve point out, balance of power politics can be understood as a repertoire of practices that states perform in order to achieve a desirable shape of the distribution of power in world politics. 12 The repertoire consists of several specific practices that states enact when they engage in balance of power politics. 13 The literature distinguishes between various forms of balancing practices such as internal and external balancing 14 or hard and soft balancing. 15 For the present purpose, it is helpful to build upon Little's distinction between 'adversarial' and 'associational' modes of balancing 16 and to conceptualise the repertoire as consisting of three types of practices: • The first typecomparative practicesconsists of those practices through which actors monitor and evaluate the distribution of power among two or more states. These practices 10 Michael Barnett and Raymond Duvall, 'Power in international politics', International Organization, 59:1 (2005), pp. 39-75. 11 Nexon, 'Balance of power', p. 335. 12 Adler and Greve, 'Security community', pp. 66-9. Practices are 'competent performances' in the sense of social activities that are endowed with intersubjective meaning (p. 66). 13 For a different understanding of repertoires, see Stacie E. Goddard, Paul K. MacDonald, and Daniel H. Nexon, 'Repertoires of statecraft: Instruments and logics of power politics', International Relations, 33:2 (2019), pp. 304-21. When the authors speak of 'repertoires of "balance of power politics"' (p. 314) in plural, they mean what we call the specific practices through which balance of power politics is enacted. However, we share their emphasis on studying variations in the combination of specific practices through which balance of power politics is performed. are underpinned by understandings of what power is and which forms of power are most relevant. With comparative practices, we mean both the institutions tasked with collecting and analysing pertinent information (such as intelligence services and general staffs), as well as the specific practices these institutions, but also other actors such as think tanks, use to analyse the distribution of power (for example, bean counts, scenarios, wargames).
• The second typeadversarial balancingcorresponds to competitive modes of balancing.
It comprises those practices through which a state (or alliance) seeks to increase its power to counter a perceived superiority of another state (or alliance). Specific practices are notably investments in military capabilities (that is, internal balancing) or the formation of alliances (that is, external balancing). The balancing dynamics are adversarial in that they result from a struggle for power in which each side strives to reach and maintain a level of power that is at least equal to that of the other side. • The third typeassociational balance makingcorresponds to cooperative modes of balancing. Associational balance making occurs when two or more competing states agree on a joint mechanism for creating and maintaining a certain distribution of power that they consider to be in balance. The most prominent specific practice is arms control, that is, agreements to circumscribe the scope and level of military capabilities. But not all associational balancing practices relate to military capabilities. The Congress of Vienna in 1814-15 and the nineteenth-century Concert of Europe were for example underpinned by a broader, and not strictly military, understanding of the balance of power. 17 Balance of power politics does not always involve all three types of practices. States can practice only adversarial forms of balancing or only associational forms. Balance of power politics can however not be enacted without comparative practices. If states do not observe and evaluate the distribution of power in terms of the presence or absence of 'balances' or 'equilibria' (however defined), then it is difficult to argue that they pursue balance of power politics. Balance of power politics presupposes that states care about differentials in power and assess and problematise the distribution of power through distinct ways of doing and saying. Put differently: A practice such as alliance formation only represents a balancing practice if the underlying objective is to redress disequilibria or maintain equilibria in a distribution of power. 18 We do not claim that comparative practices are the sole factor that determines whether and how states balance against other states. States may, for instance, perceive a power concentration to be threatening but lack the financial resources to engage in balance of power politics. Moreover, they sometimes balance against the most powerful state and sometimes ally with it. 19 Nor do we assume that states always assess the distribution of power as a whole. The balance that states seek to balance may very well be a sub-balance relating to some specific forms of power. 20 What we argue is that neither adversarial nor associational balancing practices work without some knowledge about (im)balances in the distribution of power. And this knowledge is the product of comparative practices.

Contests about balance interpretations
Realist and rationalist approaches tend to locate the three types of practices at different levels. They generally study the comparative practices as activities within states (that is, at the national level) while analysing the balancing practices and dynamics as activities between states (that is, at 17 See Paul Schroeder, The Transformation of European Politics, 1763-1848 (Oxford: Clarendon Press, 1994); Jennifer Mitzen, Power in Concert: The Nineteenth-Century Origins of Global Governance (Chicago: Chicago University Press, 2013). 18 See Adler and Greve, 'Security community', p. 68. 19 See Stephen Walt, 'Alliance formation and the balance of world power', International Security, 9:4 (1985), pp. 3-43. 20 See Steven Lobell, 'A granular theory of balancing', International Studies Quarterly, 62:3 (2018), pp. 593-605. the international level). Their study of the comparative practices focuses on the ability or inability of states to correctly asses the distribution of power. The usual benchmark is to collate the balance assessments that states made before wars with the actual outcome of these wars. 21 Research identifies several factors that account for misperceptions of the 'actual' balance, including incomplete or asymmetric information as well as individual, institutional, and political biases in the assessment process. 22 States can moreover foster misrepresentations of the distribution of power by deceiving other states about their actual capabilities. 23 These are indeed important aspects that shape how comparative practices influence balancing practices. However, realist and rationalist approaches underplay the degree to which balances are socially constructed and negotiated. What makes balance of power politics possible and seemingly imperative are practices that frame 'imbalances' in the distribution of power as threats that need to be dealt with. Besides this threat construction, highlighted by securitisation theory, 24 there is a second dimension: Comparative practices do not only evaluate balances, they also constitute them in the first place. They contribute to the very definition of what the balance isthat is, which forms of power and which actors form part of it, and in what ways. How important, for instance, is economic output compared to military capabilities? The definition of balances and sub-balances is not trivial and depends on assumptions made by the actors performing the comparative practice about the nature of power, conflict, and world politics more broadly. 25 The literature on perceptions of power has shown that states can differ in the ways in which they evaluate the distribution of power. These differences stem from the multidimensional nature of power and the variety of ways in which power and distributions of power can be conceptualised and consequently interpreted. 26 In the Cold War, for instance, the Soviet Union foregrounded military capabilities while the United States put more emphasis on economic and technical resources. 27 Because 'material distributions' are usually open to 'many interpretations', 28 the elites of states tend to foreground those dimensions of the balanceor those balancesthat best suit their portfolio of capabilities and their objectives.
There are often multiple plausible interpretations as the value of different forms of power varies both across contexts and across time and can only be known after their use. Realists and rationalists abstract from specific measures of power by looking at war outcomes, but war outcomes only work in retrospect as a benchmark for differentiating prescient from faulty interpretations. 29 And if there are no direct warssuch as during the Cold War between NATO and the Warsaw Pactthen there are also no war outcomes that could serve as benchmarks. Moreover, as constructivist and practice theory scholars argue, power is a social phenomenon that is Wohlforth, Elusive Balance, p. 303. 28 Ibid. As Booth and Wheeler, Security Dilemma, p. 61 note, 'the material facts of weapons never speak for themselves; weapons speak through the cognitive systems of their interpreters'. 29 Wohlforth, Elusive Balance, pp. 6-7. constituted by intersubjective understandings of what power is and what forms of power are valuable. 30 There are consequently no objective measures of power. This is not to say that observers and analysts are not able to differentiate between plausible and implausible interpretations of the distribution of power. The securitisation of threats is not 'separate from reality', it 'only interprets it in a particular way'. 31 Still, without objective measures of power, a spectrum of plausible interpretations remains. Comparative practices are consequently as much about getting the balance right as they are about promoting particular balance conceptions and interpretations as base for balance of power politics.
From a constructivist perspective, we therefore argue, the central question is not whether there 'actually' is a balance, but how some actors are able to make certain balance conceptions and interpretations prominent and influential in the constituencies of actors that debate, decide and enact balance of power politics. Comparative practices are performed by various actors, including politicians, government officials, military and civilian analysts, as well as journalists. If these actors differ in their balance conceptions and interpretations, disputes are likely to arise and the actors then compete over whose conceptions and interpretations resonate most in the relevant constituency of actors. The constituency of actors notably encompasses the governments and publics of the various states that are engaged in balance of power politics, but may also include other actorsfor example, peace movementsthat seek to influence these politics.
The higher the political stakes, the more likely it is that the contests are politicised. Politicisation can be understood as 'the growing salience of an issue' that involves 'a polarization of opinion about the issue' and 'an expansion of actors and audiences engaged'. 32 Balances are politicised when they are no longer debated solely among military or diplomatic experts, but become prominent and contentious political issues in a public trans-and international arena. If governments attach importance to the balance dispute, they will not only address their own publics but also a broader trans-and international audience in their quest to legitimise their balancing practices. In such an arena, states may seek to promote and substantiate their interpretations by publishing quantitative representations of the balance. The contests may morph into 'numbers wars' when the various states publicly dispute each other's quantitative representation of the balance (or the representations published by third party actors). 33

Shifts from adversarial to associational balancing
What are the dynamics of contests over balance interpretations among states? Adversarial balancingor at least its anticipationgenerally precedes associational balancing. While contests may also ariseand be resolved 34in the absence of associational balancing, shifts within the balancing practices towards associational balancing are often crucial to the dynamics of contests over balance interpretations. When competing states complement their adversarial balancing with associational balancing ( partial shift) or choose to move from adversarial to associational balancing ( full shift), they seek to make the balancing less confrontational and more cooperative. In contrast to adversarial balancing, however, associational balancing requires the development of a common understanding of the balance. For this reason, we argue, the partial or full shift to associational balancing tends to generate contests over the conception and interpretation of the balance One side could for instance assert its preferred shape of the balance by prevailing in an arms race or by winning a war.
or, if these contests already exist, make them more pronounced and intense. Three interrelated factors contribute to this effect: • In contrast to adversarial forms of balancing, associational forms require states to agree on a common 'principle of balance'. 35 Without developing an at least basic shared understanding of what the balance is and ought to be, states cannot jointly create and maintain a certain shape of the distribution of power. 36 • Each state or group of states has an interest in neutralising the relative advantages of the other side while preserving its own relative advantages. 37 Associational balance making consequently often requires and takes the form of negotiations in which each side argues for its preferred balance conception and in which concessions and compromises are needed to arrive at a balance conception that is acceptable to all sides. • In contrast to adversarial balancing practices, many associational balancing practices are based on formal agreements among the states engaged in balance of power politics. Arms control treaties are a case in point. Formal agreements increase the need for precise definitions, counting rules and quantitative comparative data. 38 They thus set higher hurdles for the development of a common understanding of the balance.
Partial shiftsthat is, situations in which the competitors move towards associational balancing but continue to practice adversarial balancingentail additional dynamics. When states pursue adversarial and associational balancing in parallel, they may play up differences within the negotiations over the conception and interpretation of the balance in order to have a public excuse for why they do not yet move past adversarial balancing. Partial shifts may last for prolonged periods of time. They last until the states involved eventually decide to fully shift from adversarial to associational balancingwith unilateral cuts as possible costly signals for their willingness to do soand make the necessary compromises to end the balance dispute. Associational balancing via arms control may take various forms. Arms control may establish equal or unequal limits on the capabilities of the competing states. Both types of limits codify a certain shape of the distribution of capabilities and hence necessitate an agreement on which capabilities form part of the balance and how these distributions ought to be distributed. 39 Moreover, in the quest to devise stable distributions of capabilities, associational balancing may involve both decreases and increases in these capabilities. Both cases are likely to fuel contests over the conception and interpretation of balances as long as they influence the present or planned force postures of the states that cooperatively strive to (re)mould the balance. Limits that are so high that they have no effects on these force posturesand do thus not involve any distributional conflictare unlikely to trigger or intensify balance disputes.
The effects are not peculiar to arms control negotiations. The Congress of Vienna, for instance, can be interpreted as a shift from the adversarial balancing of the Revolutionary and Napoleonic Wars to associational balancing as the primary mode of managing the European balance. 40 This shift involved several months of tense negotiations over how to the restructure the territorial order of Europe. The contest over the territorial balance temporarily threatened to divide the  38 Lawrence Freedman points out that 'the SALT environment' increased the demand for 'detailed comparative studies'. Lawrence Freedman, US Intelligence and the Soviet Strategic Threat (2 nd edn, Houndmills: The Macmillan Press, 1986), p. 51. 39 In asymmetrical distributions of power, the stronger states might 'persuade' the weaker states to accept unequal limits by threatening to otherwise impose the ratios by adversarial balancing. 40 For the Congress and the 'Vienna settlement', see Schroeder, Transformation, pp. 517-82.
anti-Napoleonic alliance of Austria, Great Britain, Prussia, and Russia before the four powersafter co-opting France and establishing a Statistical Commission to compile data on the territories and populations under negotiationovercame their differences and agreed on the new territorial order.
The shift towards associational balancing is therefore Janus-faced: If successful, it ends balance disputes by bringing about a shared understanding of the balance and a jointly managed balance. But the very shift from adversarial to associational balancing is likely to fuel and amplify a contest over the conception and interpretation of the balance that risks to undermine the realisation of the associational balancingwhich, in turn, may bring about a shift back to more adversarial forms of balancing.

The Cold War contests over balance interpretations
In the following, we explore the contests over balance interpretations that shaped the last two decades of the Cold War. The aim of the case study is to substantiate our argument that partial and full shifts to associational balancing shape the dynamics of contests over balance interpretations. For that purpose, we divide the history of balance of power politics in the Cold War into five phases with varying balancing practices. The five phases are summarised in Table 1. In the early Cold War, East and West engaged in adversarial balancing (phase 1). The partial shift toward associational balancing began with the Mutual and Balanced Force Reduction (MBFR) talks, which gave rise to a contest over the conventional balance in the 1970s (phase 2). The second contest, the dispute over the theatre nuclear balance, erupted when NATO adopted its Double-Track decision in 1979, which combined a planned missile deployment with an arms control offer for intermediate-range nuclear forces (INF). The so-called 'battle of the booklets', which began in parallel with the INF negotiations in 1981, marked the peak of the politicisation of the two contests (phase 3). 41 When East and West decided to fully shift from adversarial to associational balancing from the mid-1980s onward, the arms control negotiations were the forums in which they resolved their dispute over the theatre nuclear balance with the INF Treaty in 1987 (phase 4) and then their dispute over the conventional balance with the CFE Treaty in 1990 (phase 5).
Comparative practices shape balancing practices, and hence the ways distributions of capabilities are ordered, by giving political prominence and salience to particular ideas of which capabilities matter and how they are and should be distributed. At the beginning of the arms control negotiations, East and West interpreted the conventional and theatre nuclear balances differently. These differenceswhether the respective capabilities were already distributed equally (Eastern interpretation) or not (Western interpretation)fuelled the contests over the two balances. The full shift from adversarial to associational balancing was made possible by a negotiated convergence of the two sides' ideas about the balances. 42 This convergence, in turn, enabled them to order the two balances in a waythrough associational balancingthat had not been possible before. The negotiations thus underscore not only the socially constructed nature of the balances. They also changed how they were socially constructedthat is, how East and West understood the balances and ordered them. conventional capabilities but swiftly abandoned this strategy and instead settled on a strategy that relied on US nuclear superiority as a counter to the Warsaw Pact's conventional superiority. 43 The Warsaw Pact's main balancing strategy was the nuclear build-up of the Pact's sole nuclear power, the Soviet Union, which sought to end the East's nuclear inferiority and achieve a more favourable nuclear balance. 44 During these decades, there were no substantial contests over the interpretation of the conventional balance between the two alliances. Western observers widely agreed that the conventional balance strongly favoured the East. 45 The Soviet Union, in turn, announced unilateral reductions in its conventional capabilities as part of the Eastern disarmament initiatives in the UN. The arms control negotiations in the UN were, however, geared towards general and complete disarmament and not towards an associational balancing of the conventional balance between NATO and the Warsaw Pact. 46 The negotiations in the UN, which more or less ended in the mid-1960s without producing an agreement, did consequently not give rise to a contest over the conventional balance.
The conventional balance became more salient to both adversarial and associational balancing from the late 1960s onward. In anticipation of nuclear parity, NATO revised its strategy. The new strategy of flexible response put more emphasis on conventional capabilities and complemented nuclear deterrence with conventional deterrence. 47 Moreover, NATO and the Warsaw Pact signalled each other their readiness to enter into arms control negotiations over the conventional balance in Europe. Within the Western alliance, the combination of new strategy and impeding conventional arms control led to more detailed studies of the conventional balance and related reduction models. 48

Emergence: Conventional arms control and the first dispute
The MBFR talks formally began in 1973. 49 The objective was to devise a balanced distribution of military capabilities in Central Europe, which was understood to encompass the forces that the two alliances maintained in West Germany, Belgium, the Netherlands and Luxembourg on the Western side and East Germany, Poland, and Czechoslovakia on the Eastern side. The talks soon generated a dispute over the conventional balance. The Eastern and Western participations introduced proposals that diverged in their balance interpretations. 50 For the Western states, the Warsaw Pact enjoyed a conventional superiority over NATO and asymmetric reductions were consequently necessary to arrive at a balanced distribution of capabilities. For the Eastern states, in contrast, both alliances possessed approximately equal capabilities and the reductions had therefore to be symmetric to preserve the already existing balance. As the differing interpretations implied contrasting reduction models, the balance dispute came to dominate the negotiations.
Each side accentuated those aspects of the balance that it deemed to matter most, with the West framing the balance more narrowly than the East. The West proposed to concentrate the reductions on troops and tanks whereas the East insisted on reductions in all armaments, including tactical nuclear weapons that the conventional forces possessed. In December 1975, the West proposed to trade reductions in US tactical nuclear weapons for reductions in Soviet tanks in Central Europe as a way out of the dispute over which weapons were to be counted as part of the balance. 51 The East did not accept the proposal and continued to argue for the inclusion of all armaments.
Besides the armaments question, the contest was also driven by the arms control goal of equal ceilings. While the goal was primarily operationalised as an equality in manpower in the negotiations, it made a numerical balance conception dominant that also influenced the deliberations about armaments. Notably, NATO had fewer, but better tanks than the Warsaw Pact. The goal of equal ceilings, however, foregrounded quantitative inequalities while bracketing qualitative inequalities. As the Eastern adversarial balancing strategy centred on offsetting Western advantages in technology and training by a larger number of tanks, the numerical balance conception posed a problem for its argument that a conventional balance already existed. 52 The Soviets dealt with the problem by fudging the numbers to make it appear that both sides possessed approximately the same number of soldiers and tanks. As the West did not believe the Eastern numbers, this caused a 'data dispute' that deadlocked the MBFR negotiations from the late 1970s onward. While the East accused the West of playing up the data dispute in order to stall the negotiations, the West insisted that no agreement on the appropriate reductions was possible without a shared understanding of the balance.

Politicisation: NATO's Double-Track decision and the second dispute
While the balance dispute was public knowledge, MBFR did not lead to a high-profile public contest between West and East because an MBFR agreement was not a political priority for many states involved. 53  US would deploy new ballistic and cruise missiles (Pershing II and BGM-109G) to Europe (first track), and it would seek to conclude with the Soviet Union an arms control agreement limiting intermediate-range nuclear missiles in Europe (second track). 54 The statement was carefully worded to present adversarial and associational balancing as two complementary tracks, which reflected disagreements within NATO over their relative importance. 55 However, the decision practically accorded precedence to the arms control track. NATO concluded its statement by saying it would review the deployment decision should the arms control negotiations produce tangible results. 56 Moreover, the arms control track was to begin as soon as possible while the deployment of the missiles would not begin before 1983.
The decision triggered a new, second contest over balance interpretations. For NATO, the Soviet SS-20s distorted the theatre nuclear balance in Europe. For the Soviet Union and its allies, it was NATO's planned deployment that turned the balance into an imbalance. The contest soon became politicised. This politicisation was driven by peace movements and other protests, which opposed the deployment of new US missiles to Europe. 57 The ensuing public debate was characterised by the Western need to publicly defend its planned deployments, with the East seeing an opportunity for delegitimising and preventing the deployments. Yet, the peace movements did not only lobby against the deployment but also for an arms control solution, as did several European NATO members. 58 60 The booklets were widely distributed, translated into several languages, and aimed at a transnational audience.
The booklets were but one means through which the various governments sought to substantiate their balance interpretations. Other means were press statements, articles, background briefings, and interviews. Shortly before the INF negotiations, for instance, a Soviet official published an article that approached the balance in terms of launchersthat is missiles and bombersand counted 1,031 launchers for NATO and 1,055 for the Soviet Union. American officials contradicted this account. They pointed to arbitrary Soviet counting rules and argued that, if all relevant systems were included in the calculation, the Soviet Union possessed an advantage of 2,480 to 924 launchers. 61 The booklets were nonetheless special. The Western booklets played an 'important role' in the harmonisation of NATO's balance arguments, as a US INF negotiator later wrote: They provided the 'NATO nations with a common, accurate database from which to argue the Alliance's position both inside the negotiations and in the public arena'. 62 The booklets, moreover, represented the most substantial and comprehensive balance interpretations published by East and West. In contrast to the 'Soviet Military Power' series, which mostly focused on Soviet military capabilities, 63 the 'Whence the Threat to Peace' and 'Force Comparisons' series intensively discussed various dimensions of the East-West military balance. The 'Whence the Threat to Peace' editions evaluated the strategic nuclear balance between the US and the Soviet Union and the theatre nuclear, conventional and naval balances between the two alliances. 64 The 'Force Comparisons' booklets in turn analysed the three conventional military branches (ground, air, and naval forces), then assessed the relative capabilities in two European theatresthe Northern and Central region and the Southern regionand finally discussed the nuclear balance. The booklets thus merged the two contests over the conventional and theatre nuclear balances into one overall political battle over balance interpretations that went beyond these two balances.
The booklets conceptualised the conventional balance more broadly than the proposals in the MBFR negotiations did. This broader conception reflected the long-standing Eastern argument that MBFR had to encompass reductions in both troops and armaments as well as the Western MBFR argument that, while the tank imbalance was the most important imbalance, there were also imbalances regarding other major weapon systems. 65 In its booklets, the East acknowledged that the Warsaw Pact and NATO each enjoyed 'superiority in certain areas' 66 but insisted that, overall, both alliances had 'approximately equal combat capabilities'. 67 The booklet moreover claimed that the data exchanged during MBFR had confirmed that a rough balance in manpower existed, citing the Soviet data of 991,000 NATO troops vs 979,000 Warsaw Pact troops. 68 The Western booklets presented a very different assessment. According to NATO's first booklet, there was no rough equality in the number of tanks as the East claimed. 69 Rather, the Warsaw Pact enjoyed a considerably superiority in divisions (173 to 84), tanks (42,500 to 13,000), armoured personnel carriers (78,800 to 30,000) and artillery (31,500 to 10,750). 70 For the theatre nuclear balance, the booklets mirrored the diverging positions in the INF negotiations. For the West, the negotiations were about the elimination of, or alternatively equal limits for, all ground-based intermediate-range nuclear missiles such as the SS-20, the Pershing II, and Cruise Missiles deployed by the US and the Soviet Union. For the Soviet Union, this balance conception was too narrow, as it included neither medium-range nuclear capable bombers nor the weapon systems of NATO's two other nuclear powers, the UK and France. Insisting on a principle of 'equal security', the Soviet Union argued for equal ceilings for all of NATO's 'mediumrange nuclear weapon systems' on the one side and the Soviet Union's, as the Warsaw Pact's sole nuclear power, on the other side in Europe. 71 The were part of the balance. However, given their divergent understandings of what made the balance unsecure, their balance conceptions were not reconcilable. 72 Second, both sides played up the dispute to stall the negotiations and wait for the outcome of the public battle, with some US officials viewing the arms control track merely as cover for the deployments and Soviet officials hoping that the protests would force NATO to abandon the deployment. 73 Either way, the INF negotiations amplified the contest over whose balance conception and interpretation was to shape balance of power politics.
In the booklets, each side presented its balance arguments. The first edition of 'Whence the Threat to Peace' counted 986 missiles and bombers on NATO's side compared to 975 on the Soviet side. 74 It accused NATO of 'scaring the public' with a distorted balance conception by 'trying to compare the opposing forces in terms of land-based missiles only' that would ignore British and French nuclear capabilities as well US sea-based capabilities. 75 The first NATO booklet, in contrast, argued that the Warsaw Pact had 'a substantial numerical advantage' in INF and Short-Range Nuclear Forces (SNF). 76 To show that the balance was characterised by an imbalance even in the broader Soviet conception, the booklet discussed all missiles, artillery, and combat aircraft that were capable of delivering nuclear weapons. At the same time, the alliance purposively structured the discussion around the terminology of weapon systems it had developed for the negotiations. 77 Consistent with NATO's preferred balance conception, the booklet framed the Double-Track decision as relating to 'longer-range INF missile' systems. It emphasised that 'NATO presently has no missiles in this category' whereas the Soviet Union possessed SS-4, SS-5, and SS-20 missiles with 'a total of about 1,200 missile warheads'. 78 The planned US deployment of 108 Pershing II and 464 Cruise Missiles would only partially redress this imbalance and bring NATO's longer-range INF warheads to 572. 79

Depoliticisation: the INF balance
As mentioned, the Soviet Union hoped that the sustained protests would ultimately force NATO to cancel the planned deployment of US missiles to Western Europe. 80 NATO however weathered the protests and prevailed in the battle over Western public opinion. The first missiles arrived in West Germany on 23 November 1983 after the West German parliament had given its approval the day before. 81 The start of NATO's implementation of its adversarial INF balancing marked the temporary end of the associational balancing, with the Soviet Union withdrawing from the INF negotiations.
The negotiations resumed in 1985 after both the US and Soviet governments re-evaluated their balancing strategies. This was partly due to changed balancing dynamics, as nowdifferent from the first phase of negotiations in 1981 to 1983both sides were deploying INF missiles in Europe and thus engaging in adversarial balancing. It was, however, also due to the growing realisation in both governments that adversarial balancing created destabilising dynamics and that associational balancing would better serve their security interests. 82 The breakthrough in the contest over the theatre nuclear balance was in this sense enabled by developments outside the arms control negotiations, but both sides purposively chose the arms control negotiations as the pathway for resolving the contest and determining the shape of the balance.
The breakthrough happened through Soviet concessions. The new Soviet leader Michael Gorbachev step-by-step abandoned the Soviet demand for the inclusion of British and French nuclear systems, accepted the Western framing of the negotiations as being about land-based US and Soviet short-range and intermediate-range nuclear missiles and agreed to the Western 'global zero' proposal. 83 The West thus won the battle over the definition of the theatre nuclear balance. The INF Treaty was signed in December 1987 and stipulated the abolition and destruction of all land-based US and Soviet missiles with a range between 500 and 5,500 km. Put differently: The INF Treaty did not codify the balance created by adversarial balancing but rather undid the adversarial balancing.
Why did the Soviets agree to this outcome? One could, rightly, argue that Gorbachev's reform programme required an end of the resource-intensive arms race. But this does not explain why he accepted this particular balance. Informed by Western and Eastern arms control thinking, 84 Gorbachev initiated a rethinking of the Soviet understanding of what made the balance secure and unsecure. In the new balance interpretation, the arithmetic of force levels was qualified by considerations of strategic stability and common security. The last edition of 'Whence the Threat to Peace' accordingly moved past the long-time Soviet argument that the US deployments created an imbalance in medium-range nuclear forces. It noted that the US deployments had created a numerical Western superiority in medium-range forces, 85 but argued that both sides nonetheless possessed enough retaliatory nuclear capabilities to speak of 'an approximate parity'. 86 The concessions in this perspective actually improved Eastern security. While the Soviet Union had to destroy more warheads, the global zero solution made the balance even more stable by ending the arms race, enabling nuclear disarmament, and diminishing the threat of nuclear surprise attacks. 87

Resolution: the CFE balance
Two shifts in the associational balancing contributed to the resolution of the contest over the conventional balance: First, from 1987 onward, NATO and the Warsaw Pact revitalised conventional arms control through negotiations over a new arms control format. 88 They agreed to broaden the geographic scope beyond Central Europe to cover conventional capabilities from the Atlantic to the Ural Mountains, thus making all members of NATO and Warsaw Pact participants in the negotiations. Perhaps, and more importantly, both sides were now ready to discuss imbalances in weapons systemsone of the key contentious issues in MBFRand to conceptualise the balance accordingly, with the focus shifting from parity in manpower to parity in major weapon systems. Second, the successful resolution of the contest over the theatre nuclear balance both underscored the ability of associational balancing to end balance disputes and redirected the political spotlight from the theatre nuclear balance to the conventional balance, thus giving conventional arms control additional impetus. The shift to the new arms control format amounted to a restart of the balance discussion under more favourable conditions. At the end of the mandate negotiations for the new format, the two sides removed another stumbling block. In a sort of data exchange, the two sides published a further round of booklets. NATO's booklet appeared in November 1988 and the Warsaw Pact's booklet in January 1989. 89 These booklets fundamentally differed from the earlier booklets. They focused solely on the conventional balance and were meant both as public balance statements and as starting positions for the data discussion for the upcoming negotiations on a Treaty on Conventional Forces in Europe (CFE). Each side accordingly presented data for the weapon categories that it deemed relevant for these negotiations. Reflecting Gorbachev's new policy of more openness, the Eastern booklet marked the first time that the Warsaw Pact officially presented substantial data on its conventional forces, thus ending the Eastern secrecy that had hampered the discussion about the diverging balance interpretations in the MBFR talks. 90 The mandate negotiations and the two booklets made visible that the two alliances still diverged in their balance conceptions and data. In the mandate negotiations, the NATO members argued for redressing the imbalances in tanks, artillery, and armoured combat vehicles. The Warsaw Pact members, in turn, wanted to base the CFE negotiations on a broader balance conception. 91 They maintained that an overall rough parity already existed but acknowledged various imbalances in specific weapon categories. For the Warsaw Pact, the CFE negotiations had to address both imbalances that favoured the Warsaw Pact (for example, tanks, artillery, and armoured combat vehicles) and imbalances that favoured NATO (for example, combat aircraft, naval forces, and tactical nuclear weapons). For this reason, the Eastern booklet also included data on naval forces. To underscore the Soviet willingness to tackle these imbalances, Gorbachev announced substantial unilateral reductions in a speech before the UN General Assembly in December 1988.
The unilateral cut was a double signal: It showed that the Soviet Union was both willing to move past adversarial balancing and to give up its long-standing opposition to asymmetrical reductions. 92 Given the problems of the Soviet economy, one could argue that the outcome of the contest was influenced by the inabilityand, given Gorbachev's reform plans, the unwillingnessof the Soviet Union to further pursue adversarial balancing. 93 But that is only half of the story. Even if one interprets, as realists do, 94 the end of the Cold War as the management of relative decline, the emphasis has to be on management. As with the theatre nuclear balance, the contest did not end with one side acquiescing in the balance that had emerged through adversarial balancing. It ended with both sides negotiating a shared understanding of the balance and agreeing on substantial reductionsthat is, with a balance created through associational balancing.
In contrast to the INF negotiations, where the Western conception prevailed, the convergence of the balance conceptions happened through concessions made by both sides. The West notably accepted the inclusion of combat aircraft while the East abandoned its demand for the inclusion of naval forces. 95 Unlike in the mid-1970s, when the West had offered to reduce its stock of tactical nuclear weapons in return for reductions in Soviet tanks, NATO this time opposed the inclusion of nuclear weapons into the balance equation. The compromise solution was to indirectly limit tactical nuclear weapons by treating artillery and combat aircrafts that were capable of delivering nuclear weapons as part of the conventional balance. 96 The CFE Treaty, which was signed in November 1990, stipulated equal limits for five weapon categories for each side: 20,000 tanks, 30,000 armoured combat vehicles, 20,000 pieces of artillery, 6,800 combat aircraft and 2,000 attack helicopters. 97 Two levels of balance of power politics We have so far looked at contests over balance interpretations between alliances. In this section, we briefly discuss contests within alliances. We bracket what might be called 'internal' balancingthat is, how alliances cope with power differentials among their members and organise the burden sharing among themand focus on contests over the 'external' balance. 98 In order to balance collectively against an adversary, the members of an alliance have to develop and maintain a shared understanding of the balance they want to preserve or achieve and agree on the means for fostering this balance.
Balance of power politics involves two levels: The first level consists in the processes and politics through which each of the competitorsbe it states or alliancesinterpret the balance and plan their balancing strategies. The second level consists in the interplay of the balancing strategies of the competitors. 99 Both levels are implicated in the social construction of the balance. While adversarial and associational balancing are enacted at the second level, the comparative practices that underpin them are enacted both at the first level (for example, as part of the coordination process within alliances) and at the second level (for example, in the balance disputes among alliances).
At the second level, shifts to associational balancing fuel contests over balance interpretations because associational balancing requiresin contrast to adversarial balancingthat the competing states or alliances develop a shared understanding of the balance. At the first level, maintaining a shared understanding of the balance is a prerequisite within alliances for the coordination of both adversarial and associational balancing. In fact, NATO's most sustained contest over balance interpretations was triggered by a change in its adversarial balancing.
As mentioned, the strategy of flexible response that NATO adopted in 1967 put more emphasis on conventional deterrence. The US lobbied for the new strategy by arguing that conventional defence was feasible as the conventional balance had become more balanced over time. Other members, notably Great Britain and West Germany, disagreed and insisted that the balance was still so overwhelmingly in favour of the Warsaw Pact that Europe could not be defended without nuclear weapons. 100 The dispute became more pronounced after the new strategy was adopted. The alliance's first in-depth comparative study of Western and Eastern conventional capabilities, the so-called 'Relative Force Capabilities' study conducted from 1968 to 1970, served as a preparatory step for the anticipated conventional arms control negotiations, but it started as a reaction to disagreements over the implementation of the new strategy. 101 The study, though, remained inconclusive as the planned computer-based wargaming had to be abandoned due to disagreements over the underlying war scenarios. 102  alliance discussed possible MBFR reduction models in 1973, with the US arguing that NATO would be able to withstand an Eastern attack of up to 128 divisions for least 60 days while SHAPE, NATO's military command for Europe, maintained that, based on its own computer simulations, 'a collapse of a conventional-only defence would be a matter of days, at best'. 103 The alliance was nonetheless able to formulate a shared MBFR position that emphasised a conventional imbalance favouring the Warsaw Pact that had to be remedied through asymmetrical reductions. The dispute was eventually resolved with another comparative study, the so-called 'Warsaw Pact and NATO Conventional Force Capabilities' study conducted from 1973 to 1977. 104 The US, Great Britain, and West Germany acted as the three 'gaming nations' and gradually harmonised their wargaming models, which led to a convergence of their conclusions: The British and West German simulations predicted that NATO would lose the conventional war within a week, while the US simulations gave NATO 14-16 days. Despite being NATO's strongest member, it was thus not the US but Great Britain and West Germany whose interpretation of the balance prevailed.
By the time of the battle of the booklets, the alliance members broadly agreed in their interpretation of the conventional balance. It nevertheless took the alliance more than six months to produce the first 'Force Comparisons' booklet. The drafting involved several committees and rounds of feedbacks from the capitals as the booklet was meant to present NATO's official interpretation of the East-West balance in an unprecedently detailed way. The deliberations did not proceed without disputes. In March 1982, for example, NATO's International Military Staff harmonised the definitions of major weapon systems to make Western and Eastern capabilities more comparable. 105 This caused significant increases in the numbers of Eastern weapon systemsfor example, 35 per cent more armoured personnel carriers and 92 per cent more anti-tank weaponsfrom one draft to the next. West Germany protested against these changes, which led to another round of revisions, this time involving increases in the numbers of Western weapon systems, after which the alliance finally agreed to publish the booklet in early May 1982.

A transnational arena
Balance of power politics is practiced in a transnational public arena. This is not to say that all of balance of power politics takes place in public. Alliance members usually coordinate their balancing strategies in private sessions. Similarly, arms control negotiations are often conducted behind closed doors. The degree to which the participants disclose their forces and capabilities to each other is a contentious issue in many arms control negotiations. But both levels of balance of power politics have a substantial public dimension, not least because states generally seek the support of their publics as well as international society more broadly for their balancing strategies.
The transnational arena allows states to exploit the intertwinement of the two levels of balance of power politics. The booklets were not aimed at governments but at their publics and especially the Western publics. As mentioned, the Eastern strategy was to convince the Western publics of its balance interpretation in the hope that these publics would then force their governments to cancel NATO's planned missile deployments and agree to the Eastern arms control proposals (first level). This would have made the East the winner in the contests among the two alliances over the theatre nuclear and conventional balances (second level). While this strategy failed, it nonetheless underscores the dynamics that the transnational arena may add to balance of power politics. Besides states and their alliances, the transnational arena is populated by various actors. For the dynamics of balance of power politics, two groups of actors are particularly relevant: The first group consists of actors that compile and publicly disseminate comparative knowledge on the relevant balance(s). This group notably includes the International Institute for Strategic Studies (IISS) and the Stockholm International Peace Research Institute (SIPRI). The second group comprises those actors that create and seek to shape public opinion on balance of power politics. The media form part of this group, as do peace movements that protest against arms races and/or lobby for arms control. These actors often depend and draw on the data published by states, their alliances or think tanks such as the IISS and SIPRI for their interpretations of the relevant balance(s).
The first group does not only enable the second group to partake in the balance disputes. It also enables states and alliances to publicly promote certain balance interpretations without divulging their own data on these balances. In 1972, for instance, NATO published a booklet entitled 'Alliance Defense in the Seventies' in which it discussed the East-West balance based on IISS data in order not to reveal its own intelligence data. 106 In the 1980s, the West changed its strategy. Both the 'Soviet Military Power' and the 'Force Comparison' booklets drew extensively on specially declassified data. The East, in contrast, continued its tradition of revealing as little data on its forces as possible. The Eastern booklets were mainly based on already published data and explicitly mentioned the IISS and other Western think tanks as data sources. 107 As a consequence, the Western booklets contained considerably more data and more detailed balance discussions than the Eastern 'Whence the Threat to Peace' booklets. 108 At the same time, actors such as the IISS and SIPRI provide the transnational public with balance interpretations that may differ from those promoted by states and their alliances. In the Cold War, the two alliances competed not only against each other over the interpretation of the conventional and theatre nuclear balances. Because the IISS and SIPRI were widely regarded as authoritative sources for balance data and interpretations, they also had to compete with these actors. For the East-West balance, the 'Military Balance' was 'probably the most widely used reference work … with an annual distribution of 16,000 copies'. 109 While usually more in line with Western than Eastern balance interpretations, 110 the 'Military Balance' nevertheless sometimes contradicted the Western arguments. In 1979, notably, the IISS concluded that a rough equality existed in the balance of theatre nuclear forces in Europe. 111 This was at odds with NATO's argument about an imbalance in the theatre nuclear balance. The subsequent editions of the 'Military Balance' became more supportive of NATO's argument. The 1982 edition argued that 'the balance is distinctly unfavourable to NATO and is becoming more so'. 112 NATO nevertheless continued to regard the IISS as an at least potential competitor. When discussing the impact of the first 'Force Comparisons' booklet, NATO's Defence Review Committee noted that 'prior to a decision on any future edition' NATO should not only closely study the second edition of 'Whence the Threat to Peace', but also consider 'the reactions of major institutions such as the International Institute for Strategic Studies (IISS) and SIPRI'. 113 The East, in turn, treated the 106 think tanks more as authoritative sources than as competitors. The last edition of 'Whence the Threat to Peace', for instance, referred to both think tanks in its discussion of the conventional balance, noting that while 'they are not free from exaggerating data on the Warsaw Treaty forces', their analyses would show 'that the Warsaw Treaty has no superiority over NATO'. 114 On the whole, nonetheless, the contests were dominated by the interpretations advanced by the two alliances rather than those published by the IISS or SIPRI. The IISS and SIPRI enabled the transnational public to assess the balances without having to rely on the data and interpretations of the alliances. But their balance discussions mostly followed the themes set by the balancing practicesand especially the associational balancing practicesbetween the two alliances. Tellingly, despite its name, the 'Military Balance' only contained special sections devoted to balance discussions during the era of nuclear and conventional arms control from 1969 to 1991. 115

Conclusion
This article developed a constructivist approach to balance of power politics. It showed that the quest for a balance often involves contests in which states compete over the conception and interpretation of the balance. These contests are not only about whose numbers are right and whose wrong. Although Soviet officials later admitted that they fudged the numbers in the booklets to make it look like a conventional and theatre nuclear balance existed, 116 the contests were alsoand we would argue more fundamentallyabout whose conception of the balance was to underpin the remaking of the East-West military balance. In both the conventional and theatre nuclear arms control negotiations, the West argued for a more narrow balance conception than the East. The Western conception prevailed in INF whereas both sides developed a compromise conception in the CFE negotiations.
These findings suggest that research on balance of power politics has to complement its present focus on the when and how of balancing practices with a more systematic study of how imbalances and balances are constructed in the first place. As the case study showed, the making and remaking of balances is an often contentious process in which states clash over which forces and capabilities are to be counted in what ways to calculate the balances. The research has consequently to unpack how distributions of power are framed in terms of 'balances' and 'imbalances' and how some of these interpretations come to shape, both domestically and internationally, the balance of power politics that states pursue.
The findings moreover underscore that associational balancing tends to initially intensify the contests before it potentially contributes to their resolution. Our case study centred on arms control producing balances structured by equal limits. In order to better understand the scope conditions under which associational balancing fuels contests over balance conceptions and interpretations, further research could look at cases that involve balances with unequal limits (for example, naval arms control in the interwar period) and cases of associational balancing that were not based on arms control (for example, the nineteenth-century Vienna settlement and order).
What does the case study imply for the present phase of renewed great power competition which, so far, involves adversarial but no associational balancing among all the great powers involved in the competition? Balance disputes may arise before the great powers begin arms control negotiations. The Cold War experience nonetheless suggests that the initiation of such negotiationswhile usually thought of as a de-escalation of great power competitionwould make these balance disputes more intense and politicised.