‘We need allies and collective security’: How the end of the Second World War caused multilateralism in Europe and bilateralism in Asia

Abstract

Why is there no NATO in Asia? Literature on this question is selective and incomplete. This paper develops a new theory with determinate predictions regarding patron and clients’ alliance design preferences, the alliances that result, the commitments therein, and alliance duration. A subtle but nonetheless persistent form of entrapment problem exists with clients that don’t want a war yet fear adversary aggression. Clients’ commitment to collective security is a hand-tying costly signal that assures the patron of client resolve to defend the status quo and reduces the probability and costs of entrapment. Patrons will rationally prefer to join in an alliance clients who have already made a collective security commitment. Clients are more likely, the paper shows, to make such commitments when their adversary credibly threatens to militarily occupy at least one of them. This is more likely in land than in sea theatres. When clients fail to realise collective security, patron efforts to impose it on them will fail, will result in short lived multilateralism, and will force bilateralism on the patron. The paper uses new archival evidence from Britain and Australia to show how this strategic framework explains variation in alliance design in Europe and Asia.

Why is there no NATO in Asia? Under what conditions might a multilateral alliance between the United States and its allies in Asia and/or the Indo-Pacific emerge? Between 1949 and 1955, the United States committed to 7 different defensive alliances with 22 partners in Europe and Asia. Why were 14 European clients incorporated into one long-lasting multilateral alliance while 8 Asian clients scored 5 mostly bilateral alliances and a multilateral alliance (SEATO) that did not last? These alliances have been central to US grand strategy and world order since the beginning of the Cold War; this question has substantial scholarly and policy implications. Our knowledge of why multilateralism took root in Europe and bilateralism in Asia however remains selective and incomplete.

This paper provides theory and evidence to show that the source of regional variation in alliance design is the US European clients’ prior defensive commitment to each other through a collective security system.¹ Such a prior commitment was absent in Asia, and these client commitments to each other will, the analysis demonstrates, most assure patrons contemplating an alliance regarding a fundamental entrapment problem. John Foster Dulles himself, the architect of early US Cold War alliances, explained in January 1954 that ‘we need allies and collective security’ and that ‘we rely principally on a community security system’ that is ‘well equipped to punish any who break in and steal’.² When patrons refrain from offering a defensive alliance commitment to their clients until the clients have made such a commitment to each other, a durable multilateral alliance between the patron and clients will result. But such client commitments to their own security is so reassuring that patrons might be tempted to force such multilateralism on their clients. Under these conditions, multilateralism will fail, and the patron will have to settle for bilateral alliances. The paper also shows that the Soviet threat of credible military occupation of US clients in Europe but not Asia – despite Soviet leader Joseph Stalin’s military mobilisation for the occupation of northern Hokkaido – explains why US clients in Europe but not Asia committed to each other’s security. The theory and evidence show that if China today generates a threat of credible military occupation of US Asian/Indo-Pacific allies, an Asian NATO will likely result.

The paper synthesises balance of threat, geography, and alliance abandonment and entrapment theories to develop a new theoretical framework that explains continuity in patron alliance design preferences, variation in client alliance design preferences and the alliances that emerge.³ The threat of the credible adversary occupation of clients given the balance of offensive power and geography overcomes the clients’ collective action problem and, through aligning patron and client threat perceptions, overcomes the patron’s entrapment problem whereby patrons perceive different threats from an adversary than their clients. The empirical analysis tests this theory using extensive new archival evidence from Britain and Australia as well as Foreign Relations of the United States sources. The Soviet occupation of East Germany and East Berlin instilled rational fear of military occupation in territorially contiguous US European clients in the late 1940s, which the analysis shows incentivised them to commit to each other’s security in a system of collective defence to deter Soviet challenges. This assured the Truman administration and incentivised US President Harry Truman to similarly commit to this nascent European collective security system.

The sea-based geography of the Asian theatre is insufficient to explain why bilateral and not multilateral alliances emerged there. The region’s maritime geography cannot explain why Stalin authorised a large military mobilisation of the half of northern Hokkaido in 1945 only to call it off in the face of US opposition. If geography is determinative, why did the Soviet leader authorise such a mobilisation in the first place? A Soviet occupation of northern Hokkaido would likely have seen Japan come to have the same divided status as Germany. Had Stalin occupied northern Hokkaido, US clients in Asia would have perceived greater threat and been more likely to commit to each other and eventually US-occupied Japan in a multilateral defensive alliance. The Soviet occupation of the volcano-laden Kurile Islands and mountainous southern Sakhalin but non-occupation of Japan, however, left US clients in the region – Canberra, Wellington, and Manila – less threatened by potential Soviet revisionism emanating from Japan and less committed to each other’s security. This meant that US plans for multilateralism – first canvassed through the region by Dulles in 1951 and then again imposed through SEATO in 1954 – either did not fly or did not last. In short, to understand variation in the design and nature of the US commitment to its clients in Europe and Asia in the early Cold War, we need to look back to the geopolitical status quo at the end of the Second World War and the client commitments to each other that resulted. The fighting in and resolution of the Second World War have tended to be studied separately to the onset of the early Cold War; this has limited our ability to better understand regional variation in early Cold War alliance design.

The findings here have several implications for scholarship and policymaking. They show that a common strategic logic regarding patrons, clients, collective security, an adversary, and the latter’s credible threat of military occupation explains multilateralism in Europe and bilateralism in Asia. Hemmer and Katzenstein claimed in an important contribution that ‘a direct line from a certain type of threat (cross-border Soviet attack) to a particular institutional form (multilateralism) cannot be drawn in Europe’.⁴ The argument and evidence presented here offer such a direct line. The paper provides further evidence of the rational incentives for patrons to pursue and encourage collective security and multilateralism in defensive alliances with their clients.⁵ The theory and evidence also suggest that credible threats of military occupation can be a double-edged sword: greater power, influence, or strategic depth must be weighed against how this incentivises neighbouring states to ally together. From a policymaking perspective, the paper shows that if Chinese military systems and/or behaviour in the Indo-Pacific lead policymakers in states like Australia, Japan, and the Philippines to perceive a more credible threat of Chinese military occupation, current institutions like the quadrilateral security dialogue (QUAD) may come to look more like a multilateral defensive alliance.

The remainder of this paper reviews the literature on variation in alliance design in Europe and Asia then justifies the case selection of the Dunkirk, Brussels, and NATO alliances in Europe and US alliances with Japan, the Philippines, Australia, and New Zealand in Asia regarding questions of alliance design and the ‘why no NATO in Asia’ question. The paper then develops a theory of multilateral alliance design that has determinate predictions regarding the type of generic alliance desired by patrons, clients, their commitments, and the duration of the resulting alliance. The empirical analysis uses new archival documents to test this theory on European collective security in 1947 and 1948, the US commitment to this system in 1949, US commitments to Australia, New Zealand, Japan, and the Philippines in 1951, and, briefly, SEATO in 1954.

Literature: Alliance design in Europe and Asia

Literature on the institutional variation of US alliances in Europe and Asia is indeterminate and incomplete. There are four prominent sets of arguments. The first set focuses on hegemonic control. These arguments link greater power to a desire for multilateralism but cannot explain the link between the same preponderant US power and bilateralism elsewhere.⁶ For Ikenberry, multilateralism reduced transaction costs and helped facilitate ‘coordination that incentivised other states into desirable and stable policy orientations’.⁷ But if multilateralism corrected potential institutional inefficiencies, why was it not realised in Asia to better shore up US allies in the Korean War and better integrate Japan into the emerging regional and global order? For Victor Cha, bilateralism allows patrons greater control over clients.⁸ This nicely explains why the US alliances with South Korea and Taiwan were bilateral; such control was necessary to ensure that Seoul and Taipei didn’t drag Eisenhower into a(nother) war with China. Is bilateralism optimal if patrons don’t believe that their client is trying to drag them into a war? Cha’s framework whereby a patron exerts tight control over a single ally – US planners did not perceive that Tokyo wanted to drag them into a war but nonetheless exerted high control – will tend to be the exception rather than the rule. Indeed, Cha’s theoretical set-up of a patron and one client makes multilateralism logically impossible. It is also at odds with Washington’s desire for a multilateral alliance with its European and Asian clients.⁹ The theory developed here reorients attention to the patron’s clients to explain regional variation in alliance design. Izumikawa used network analysis on US clients in Asia to explain bilateral alliances there despite the US preference for multilateralism but did not address Europe.¹⁰ Moreover, Izumikawa’s theoretical model is restricted to the patron and its clients and allies such that the adversary cannot influence the alliances that emerge. The model developed here allows adversaries to influence the alliance designed to deter them.

The second argument for variation in alliance design in Europe and Asia focused on ideational variables and, specifically, US racism.¹¹ Asian collective identities were, according to this argument, considered inherently inferior to those of the United States and Western Europe. Common religion, democratic values, trust, and other affinities led to shared values and commitments and multilateralism in Europe, but racism towards Asians led to bilateralism in Asia.¹² Several policies, however, could have been logically consistent with these racist beliefs. Should such racist collective identities lead us to expect the United States to have ignored the region, militarily occupied the region, or sought a military alliance with its members? Hemmer and Katzenstein’s argument cannot explain the post–Korean War consistent US preference for an Asian NATO. Endogenous to US racism, the analysis here will show, was a common strategic outlook that applied to Europe and Asia which extant literature has thus far neglected. Another prominent argument in this tradition points to local – for example, Indonesian and Burmese – disinterest in such an arrangement.¹³ This, however, doesn’t explain the push by the Filipinos, Thais, South Koreans, Taiwanese, and others for membership in a regional defensive alliance. Other arguments have a greater role for different ideational variables including hierarchical social contracts, religious, racial, socialist, and democratic social logics, and associated norms, ideas, identities and regional institutional variables but do not attempt a comparative analysis of alliance variation in Europe and Asia.¹⁴

The third argument for regional variation in alliance design posited that the answer lay in social psychology and prospect theory. For He and Feng, high/low levels of threat are framed in the domain of losses/gains, the variation between the best and worst outcomes in a bilateral alliance is smaller than that in a multilateral alliance, making multilateral alliances riskier than bilateral ones, and greater threat perception drives multilateralism while lower threat perception drives bilateralism.¹⁵ These assumptions explain the US persistence with multilateral alliances in the early Cold War. They also explain why Washington’s SEATO commitments were less than its NATO ones. But the claim that multilateral alliances are riskier than bilateral ones runs into problems; the costs of alliances for great powers are not only constraints but also entrapment. Did the Eisenhower administration view the Taiwan Straits and Korean peninsula in the early 1950s as less risky than Europe? Must high-level threats be framed in the domain of losses? Our knowledge of reference points and framing is limited. As Jack Levy pointed out, prospect theory ‘is a reference-dependent theory without a theory of the reference point’.¹⁶ Yet He and Feng’s model of alliance variation in Europe and Asia depends on these assumptions.

The fourth set of arguments focuses on geography but offers indeterminate predictions regarding variation in alliance design.¹⁷ Scholars like Mearsheimer and Levy and Thompson have made much of the stopping power of water; one central implication of this literature is that balancing alliances are much more likely to form in land theatres than in maritime theatres.¹⁸ But this says nothing about whether any alliances that result will be multilateral or bilateral. Without further specifications, this literature at most predicts that alliances are more likely in Europe than Asia and cannot explain any alliances, let alone bilateral ones, in Asia. Literature on alliance creation offers little to remedy this. For Glenn Snyder, who allies with whom results from ‘a bargaining process that is theoretically indeterminate’.¹⁹ Walt’s Balance of Threat theory is also indeterminate insofar as it does not clarify what combinations of offensive power, geography, and perceived aggressive intention give rise to what types of alliances.²⁰ Mearsheimer, Levy and Thompson, Snyder, and Walt might together suggest that alliances are more likely in Europe than Asia, but the Korean War and resulting US alliances in Asia make even this conclusion problematic. These influential arguments cannot make determinate predictions regarding alliance design in Europe and Asia without further specifications that the model developed here provides.

We lack a theoretical model of a patron, multiple clients, and an adversary that explains patron and client preferences for alliance design, multilateralism in Europe, and bilateralism in Asia. The theoretical model developed here fills that gap. It aims to move beyond extant literature by developing a dynamic theoretical model of a patron, clients, and adversary that offers determinate predictions regarding (1) patron alliance-design preferences, (2) clients’ alliance-design preferences, (3) whether a multilateral or bilateral alliance results, (4) the extent of the security commitments therein, and (5) how long-lived the resulting alliances will be, which explains the origins of alliance variation in Europe and Asia. Grasping the core dilemma facing a patron that considers extending a defensive alliance to its clients confronted by a common adversary offers a solution.

Theory: Patrons, clients, alliances, and collective security

Patrons can evade most military commitments within an alliance but nonetheless tend to be most wary of entrapment – defined as ‘being dragged into a conflict over an ally’s interests that one does not share, or shares only partially’ – when signing up to defensive alliances with their clients.²¹ Committing to military alliances is essentially drawing a red line that can signal to an adversary how far they can challenge without retaliation.²² Great powers who by definition have many global interests and commitments would rationally worry that adversaries and neutrals would conclude that an alliance with one/some clients but not others would leave the latter ripe for the picking. Adversaries may also respond to the alliance by developing alliances with their clients and pressuring neutrals to pick sides. Moreover, even if patrons can find a way to not have to fulfil military commitments within an alliance, they would worry that every evaded chance might earn them a reputation for irresolution in that alliance or elsewhere.²³ Yarhi-Milo, Lanoszka, and Cooper thus argued that if patrons are to offer defensive alliances rather than arms to their client(s), they will tend to believe that they and the client(s) share convergent threat perceptions.²⁴

If patrons decide that a defensive alliance with one or more clients is worth this spate of hazards, the first question they must ask is whether their client/s is/are more likely to challenge a common adversary than the adversary is to challenge the client/s. This is a complicated question: the clients have incentives to convince the patron that they want peace and their adversary wants war, while the adversary, not wanting the alliance to form, wants the patron to believe the opposite. Also important are the clients’ military capabilities and skills vis-à-vis the adversary: if the patron came to their aid in a conflict, would the patron have to shoulder most of the military burden? Client and adversary incentives to deceive here are high, greatly complicating the patron’s assessment. The answers to these questions and incentives to deceive can also change with the balance of power, resolve, and information in ways that are hard for patrons and scholars to predict.²⁵

A patron may decide that the clients’ probability of provoking a war with the adversary is greater than or about equal to that of the common adversary. Under these conditions, the patron might offer no alliance, concluding that any military commitment would embolden the client to challenge the adversary, or offer a bilateral alliance which bestows on the patron significant control over the client’s military power.²⁶ What if the patron concludes that the adversary will be more likely to make a military challenge than the clients? Here, the patron must ask other questions. If the patron fights for the clients, how good a fight will the clients offer against the common adversary? Is client military interoperability with the patron high?²⁷ If there is more than one client, another way of asking this is assessing how willing and capable the clients are of fighting for and with each other against one or more common adversaries.²⁸ Clients here face a daunting reassurance problem: how can they reassure the patron that their resolve and capabilities to defend the status quo are high? A commitment to fight if a state’s own security is threatened might be considered a baseline for a patron to consider offering a state an alliance. Clients have incentives to exaggerate their capabilities and resolve to fight the adversary. A formal commitment to fight if other clients are attacked signals greater resolve to defend against and deter adversary aggression through aggregating resources and sharing burdens. All else equal, better to have to fight for clients that have already committed to each other’s defence. This collective security assures patrons that the probability and/or costs of entrapment are reduced. Patrons facing the trade-offs of offering alliances to status-quo-oriented clients will tend to find this assurance regarding entrapment challenge their core problem. Client collective security will be an assuring signal for racist and non-racist patrons, be equally important to the patron regarding in-group and out-group clients, and assure patrons who view clients in gains and losses frames. Client collective security commitments will tend to be such a strong costly signal that patrons will optimally condition the provision of a defensive alliance on such commitments. Such client costly signals should therefore cause long-lasting alliances with patrons.

H1: patrons will want their (status-quo-oriented) clients to commit to their own security before the patron does

H2: alliances where patrons join clients’ collective security alliances will tend to be long-lasting

If clients struggle and/or fail to achieve a collective security commitment, the assurance that such commitments offer regarding entrapment may lure patrons into manufacturing or imposing such commitments on their clients before they commit to an alliance with them. Resulting multilateral alliances will tend to not last due to the clients’ weak or non-existent commitment to each other. In these cases, patrons will have to fall back on bilateral or, if two clients have already committed to each other, trilateral alliances.

H3: alliances where patrons manufacture or impose collective security commitments on their clients will not last

H4: where manufactured or imposed collective security fails, patrons will have to settle for bilateral or trilateral alliances

If a patron’s clients have committed to a system of collective security, a patron defensive commitment to only some of these clients might weaken the former collective security system through encouraging a challenge against those clients without a commitment from the patron. A commitment to consult only where the clients have committed to fight for each other may signal disinterest in the clients to an adversary. If a patron commits to a group of states that have already made a collective security commitment, the patron’s commitments will therefore be qualitatively similar to fight if any of the group are challenged.²⁹ If two clients have committed to fight for each other, the patron would likely also commit to do so when joining such an alliance.

H5: If a patron makes an alliance commitment to a group of clients, the patron will tend to make the same commitment that the clients have made to each other

What would influence whether a patron’s clients have committed to each other’s defence? The most important variable will likely be the threat posed by an adversary given the balance of offensive power and geography. Most explanations of variation in alliance design in Europe and Asia have neglected the adversary. However, if an adversary has militarily occupied one of the clients or, given geography and the balance of offensive military power, could easily do so, many if not most clients will feel threatened. If conquering one client would make it easier for the adversary to conquer more or all of them, all clients will have incentives to commit to each other’s defence. This would force an adversary to face the prospect of fighting all if it fights one and reduce the probability that the adversary could pick them off individually. Collective security signals to the adversary that the latter cannot pick the clients off at will but must fight them together. However, if geography and the balance of power leave the clients perceiving less threat by potential adversary aggression, they will be less likely to commit to each other’s defence. Incentives to free-ride on the efforts of others will be higher because the threat of military occupation or attack is more remote, and the lack of a common threat may allow previous rivalries and fears amongs the clients – perhaps towards each other – to be more salient.

H6: a patron’s clients will be more likely to have committed to collective security if a common adversary can, given the balance of offensive power and geography, credibly threaten to militarily occupy them

The causal mechanism whereby the adversary’s credible threat of military occupation of any of the clients incentivises the clients to commit to each other’s defence could also involve perceptions of change in adversary intentions. Changes in the ease of further military occupation given the new balance of power and geography are, however, likely to be at least as if not more important than changes in the balance of information.³⁰

Land theatres are more likely to allow adversaries to credibly threaten military occupation of clients than naval theatres. In land theatres, sufficiently high military power and the military occupation of even part of a state can open the door to the rapid occupation of the remainder of that state and its territorially contiguous neighbours. Maritime theatres require greater amounts of military power and expertise to post the same credible threats of military occupation. Quicker blitzkriegs are more likely in land theatres. In naval theatres, the occupation of part of an important state can allow a challenger to occupy all of that state and develop or acquire naval forces to credibly threaten others in that maritime theatre with military occupation, but this will tend to take more time and greater resources. It is not the case, however, that maritime theatres are more likely to lead to bilateral alliances. Maritime theatres should, all else equal, reduce threats of credible military occupation, which will reduce the probability of any alliances forming.³¹ If a threat emerges in a maritime theatre, geography alone is indeterminate regarding resulting alliance patterns.

H7: adversaries will be more likely to credibly threaten military occupation of clients in land-based than in naval theatres

The adversary credible threat of military occupation of the patron’s clients is more likely to cause the threat perceptions of the clients to converge and be more inclined to commit to each other’s defence absent the patron. It also aligns client threat perceptions with the patron, which then reduces the patron’s moral hazard concerns. To summarise, incorporating patrons, clients, adversaries, credible military occupation threats, and collective security within a model of alliance formation offers determinate predictions for alliance design, commitments, and duration.

To test the above theory of alliance design, the relevant universe of cases are all asymmetric interstate defensive alliances formed between great powers and their independent clients during the multipolar, bipolar, and unipolar eras. Morrow has shown that these alliances involve autonomy and security trade-offs; security trade-offs for patrons can include clients’ and allies’ assurances regarding the costs and probability of entrapment.³² To explain the absence of an Asian NATO, this paper conducts a comparative analysis of the origins of the 1947 Dunkirk alliance between Britain and France, the 1948 Brussels treaty between Britain, France, Belgium, the Netherlands, and Luxembourg, the 1949 NATO alliance, the US 1951 preference for a multilateral alliance with Japan, the Philippines, Australia, and New Zealand, the trilateral and bilateral alliances with these clients that resulted, and the 1954 SEATO alliance. Literature on variation in alliance design in Europe and Asia has not addressed the two European alliances that preceded the NATO alliance despite their theoretical relevance. The paper uses process tracing and ‘causal process observations’ to identify, measure, and locate the sources of patron concerns with defensive alliances and client fear of military occupation.³³ I searched the National Archives in Britain, the National Archives and National Library in Australia, the National Archives and Records Administration and Truman Presidential Library in the United States, as well as several online repositories, for evidence of client fear of adversary military occupation, how this influenced their orientation to collective security and alliances, and patron concern about alliance entrapment and how this influenced their approach to an alliance. I also searched for evidence regarding variables outlined by alternative explanations that link great power desires for control, institutional efficiencies, and racism to alliance design.

Note: Assumes patron believes adversary desires war with clients more than clients desire war with adversary

Figure 1. Military occupation and collective security.

Europe: Credible Soviet threat and collective security

This section tests H6 and shows that the waxing and waning of the Soviet threat drove London and Paris and their Benelux partners to converge on the threat from Moscow, commit to each other’s security, and overcome differences regarding Germany in the absence of a US security commitment. A British memo from January 1947 noted the utility of an alliance with France but doubted that ‘an Anglo-French alliance can be concluded without a prior settlement of our differences of opinion about Germany’.³⁴ Against models of hegemonic control, this was first and foremost a British initiative; there does not seem to have been US pressure or control behind the scenes. Much British thinking directed the purpose of the treaty with Paris against ‘a recrudescence of the German menace’.³⁵ But for the British and to a large extent the French leadership, the Soviet threat lurked behind this.³⁶ As a January 1947 memo remarked, ‘any suggestion that the Treaty was not directed primarily against Germany would rouse suspicion and perhaps opposition not only in Moscow but in some other quarters’.³⁷ A memo regarding a potential British–French alliance wondered ‘how real are the risks vis-a-vis the Russians’.³⁸ The Chiefs of Staff Committee were in favour of inserting a collective defence clause into the alliance with Paris but, anticipating Soviet objections, noted that ‘it might have to be done “off the record” and unobtrusively and not in virtue of any specific clause in the Alliance’.³⁹ The French decided against a final proposed agreement because its annexes could not remain secret and ‘would make it impossible to say that the Alliance was only concerned with Germany’.⁴⁰ As H6 expects, the Soviet threat drove Britain and France’s commitment to each other. A 27 February treaty draft closely mirroring the signed Dunkirk alliance eventually committed London and Paris to consult with each other, Washington, and Moscow ‘where appropriate’ to address (Soviet) threats emanating from Germany.⁴¹ The beginnings of European collective security were thus born in February 1947 with most stops and starts determined by anticipated Soviet reactions. As Baylis noted, against the expectations of models of hegemonic control, ‘there was not a widespread consensus in the Foreign Office or in British political circles generally, that the ultimate objective should be … a wider security alliance involving the United States’.⁴²

Regarding the Belgians and Dutch, who seemed eager to join the party, a 3 March memo advised that, given Soviet reactions, ‘it would be better to go slow’.⁴³ By May, however, Bevin had decided that ‘the conclusion of such treaties (with Belgium and the Netherlands) would probably be desirable … the time has come when we should start to bind our friends in Western Europe more closely to our side’.⁴⁴ For the Chiefs of Staff Committee Joint Planning Staff in June 1947, the ultimate goal was a ‘closely knit Western European Association in alliance with ourselves’.⁴⁵ H6 is supported by the Soviet military threat being the chief motivator of these commitments. The proposed alliance was designed to stand up to ‘land attack from the east … we should welcome any arrangement which tends to bind these States to us and offers us an opportunity of co-ordinating defensive arrangements between them’.⁴⁶ Only the Soviet Union posed such a threat, and Stalin’s forces in East Germany and East Berlin made this threat more menacing. By December 1947, Bevin insisted to Bidault that British and French military talks ‘must be begun quite soon, but in the most confidential manner, and we should try to bring the Americans in later’.⁴⁷ The foreign secretary would write a memorandum to the Americans in January 1948 warning that ‘we shall be hard put to stern the further encroachment of the Soviet tide … we in Britain can no longer stand outside Europe and insist that our problems and position are quite separate from those of our European neighbours’.⁴⁸ It was the Soviet threat that prompted Bevin to look to collective security.

The Americans were well aware that while much was made of Germany, the real concern, as H6 expects, was the Soviet Union. For director of European Affairs John Hickerson, ‘extension of the Dunkirk Pact against German aggression is highly dubious’.⁴⁹ Thus, Marc Trachtenberg pointed out that ‘the Soviet threat was the overriding concern … these texts were directed at Germany for essentially tactical reasons’.⁵⁰ The German threat was, in Bidault’s words, a ‘convenient myth’.⁵¹ The executive secretary of the NSC provided further support for H6 when he reported that the driving force of European collective security was ‘fear of Soviet-Communist aggression’.⁵² A British memo from 1948 noted that growth in Western Union military strength ‘will constitute a progressively more effective and firmer deterrent to Russian aggression’.⁵³ A February 1948 British memo recommending participation in the Brussels treaty identified the source of threat as ‘Germany or any other state which might unite with Germany directly or in any other form’.⁵⁴ A July 1948 Five Power Military Committee report called for the focus of the ‘Strategic Concept’ to be ‘a definition of Russian strategic aims’ and called for ‘an estimate in the broadest terms of the total forces required to defeat Russian intentions’.⁵⁵ Another British memorandum was even more explicit: ‘no major power except Russia at present constitutes a threat to our interests’.⁵⁶ Most of London’s French and Benelux interlocutors would have privately agreed with this. The State Department record of the Washington Exploratory Conversations on Security between the United States and its European clients noted, in a manner that the Europeans would have fully agreed with and which strongly supports H6, that:

The westward expansion of Soviet power since the defeat of Hitler has rendered the Soviet Union strategically capable at the present time of dominating the continent of Europe by force. Soviet forces are so grouped and organized that they could take the initiative in military action at short notice.⁵⁷

Bevin would often speak of the need to ‘organise and consolidate the ethical and spiritual forces inherent in this Western civilisation’ and explicitly call for ‘some form of union in Western Europe, whether of a formal or informal character, backed by the Americans and the Dominions’.⁵⁸ But behind all such talk of moral or spiritual forces that smells of racism, the Soviet threat drove Western European collective security thinking.

By 19 February, the Benelux countries were all in on collective defence: ‘in the opinion of the three governments, this mutual aid in case of armed aggression, authorised by Article 51, must be automatic and immediate’.⁵⁹ The Soviet-sponsored coup in Czechoslovakia convinced the Europeans and especially the French of the need for collective security in the face of the Soviet threat.⁶⁰ The US ambassador to France wrote to Marshall that ‘what he (Bidault) really wants more than anything else, is a concrete military alliance with definite promises to do definite things under certain circumstances’.⁶¹ Although the British embassy warned that an imminent Soviet alliance proposal to Norway was partly motivated by a desire to drive a wedge in West European collective security efforts, Bevin declined to offer Oslo a participatory role in the collective security discussions because ‘France and the United Kingdom, with the Benelux countries, could not by themselves effectively defend Scandinavia against pressure’.⁶² Moreover, the British defence chiefs were now coming around to the position that ‘if Europe was overwhelmed (by Soviet aggression), the United Kingdom would be threatened as never before and might well not survive’.⁶³ This provides striking support of the Soviet threat driving the European commitment to collective security, as H6 expects.

The Soviet threat to Europe had incentivised US partners there to externally balance and ‘afford the party so attacked all the military and other aid and assistance in their power’. It would take one day for Bidault to advocate ‘establishing common military technique and standardization arms among five, bearing in mind US technique and equipment’.⁶⁴ Western European collective security was driven by the Soviet threat to Western Europe and exacerbated by the coup in Czechoslovakia and threats to Finland and Norway.⁶⁵ As Rosato concurred, the Soviet ‘potential to dominate’ motivated European collective security.⁶⁶

Europe: The US multilateral commitment

This section tests H1 and shows that the US commitment to the Western Europeans was driven by the assurance that European collective security provided. This was reliable evidence of European clients’ resolve to defend each other to deter Soviet challenges and evidence that their threat perceptions of the Soviet Union had sufficiently converged. That the Europeans themselves were aware that their commitment to each other would assure the Americans provides strong support for H1. Against models that emphasise civilisational or racist influences, it was a strategic rationale that drove the US commitment. As Odd Arne Westad pointed out, ‘the setting up of NATO was not about a civilizational definition of a European core’.⁶⁷ Nor was it about a power play; the fundamental US goal was to seek reassurance regarding the Europeans’ commitment to stand up to the Soviet threat that collective security provided.

The British and Europeans convinced Marshall that they were sufficiently committed to their own collective defence.⁶⁸ He informed President Truman, as H1 expects, that the talks in Brussels would be an index of European commitment to collective security and ‘provide a starting point for our consultations with them’.⁶⁹ As the US charge in Belgium wrote to Marshall, the Europeans

all feel that US Government quite naturally expects European nations do utmost on their part and they only hope that when US Government convinced they have done so, US would be disposed to lend support.⁷⁰

European initiative towards collective security was, as H1 expects, highly assuring to Washington. George Kennan, for example, had advised Marshall that

the initiative must come from Europe, and the project must be worked out over there … if they develop it and make it work, there will be no real question as to our long-term relationship with it, even with respect to the military guarantee. This will follow logically from the circumstances.⁷¹

Marshall replied to the British ambassador on 20 January that ‘I wish to see the United States do everything which it properly can in assisting the European nations in bringing along this line to fruition’.⁷² The American was, however, leaving the ball in Bevin’s court.⁷³ Indeed, Bevin’s ‘spiritual federation’ for European security at the end of 1947 remained ‘vague and ill-defined’.⁷⁴ Hickerson reiterated to the British ambassador, consistently with H1, that while ‘any adequate regional defense system for western European countries should envisage defense measures to be taken in the event of aggression or attack from any source … if the European nations created such an organisation and made it work, there would be no difficulties in settling our long range relationship with it’.⁷⁵ As Hickerson clarified, ‘every proof that the free states of Western Europe could give that they were resolved and able to stand on their own feet’ would elicit a firmer US commitment.⁷⁶ This provides strong support for H1. The British ambassador understood that US involvement would ‘be at the second stage rather than during the initial phases of the scheme’.⁷⁷

The British begged and pleaded for a prior US commitment, but, as H1 expects, the Americans were unmoved. The British ambassador informed the under secretary of state that European nations would commit more and earlier if Lovett offered ‘some indication of the extent of United States support of the program, particularly if they were to commit themselves to any defense system’.⁷⁸ The American, however, did not budge and pointed out that ‘the British government are not yet in a position to give firm assurances as to the role Britain intends to play in operations on the continent of Europe’.⁷⁹ The British, however, persisted. Inverchapel begged for anything resembling a US commitment: ‘it hampers Mr. Bevin not to know … what arrangements he could propose … that they would commend themselves to Mr. Marshall from the point of view of possible United States participation at a later stage’.⁸⁰ But as Lovett replied, in a manner that confirms H1 and the assuring value of local initiative, ‘you are in effect asking us to pour concrete before we see the blueprints … it is very important that the initiative in this matter remain in Europe’.⁸¹ US insistence on European initiative in the face of British persistence for some US commitment is found throughout the 1948 diplomatic record and provides strong support for H1.⁸² Against models of hegemonic control that point to the United States as the source of multilateralism in Europe, the diplomatic record leaves no doubt that the Brussels treaty was a European initiative. The Americans looked on favourably but rationally stuck to their guns that collective security in Europe had to be a European project and do not seem to have applied pressure to realise this outcome other than the vague promise of their full support if the Europeans could pull it off. Marshall informed his ambassador in Italy, viewed as vulnerable to ‘further fifth-column action along Czech model’, that ‘form and extent of our relationship to them (European collective security discussions) cannot be determined until we can study final product Brussels talks’.⁸³ As H1 expects, the Truman administration were holding out for European assurance of their commitment to each other.

At the US–UK–Canada meeting on 22 March, US ambassador Douglas flagged three options: an extension of the Brussels Pact, an Atlantic Pact, which might include other regional pacts such as a Mediterranean Pact, and a worldwide self-defence pact based on Article 51 of the United Nations Charter.⁸⁴ Although a Policy Planning staff memo noted that any US assurance should, consistently with H1, ‘leave maximum freedom of method compatible with effective assurance of reciprocal support from them’. (Western European clients), it recommended an immediate diplomatic assurance that in light of West European collective security commitments, if these were upheld, ‘the U.S. will consider armed attack against them to constitute armed attack against the U.S.’, including a ‘willingness to consider whether a given case of indirect aggression should be deemed armed attack’, co-ordinate ‘military, informational and other anti-communist efforts’, and strengthen West European military power.⁸⁵ The indigenous European commitment to collective security had assured the Truman administration and was, as H5 expects, driving a similar US commitment. Truman and Acheson do not seem to have wanted control over their West European clients as hegemonic control theory expects; the Europeans’ collective security commitments put the onus on the US commitment to mirror this.

US assurances were to be offered ‘on a basis of reciprocal military undertakings which would predicate resolute action on their part … but leave maximum freedom of method compatible with effective assurance of reciprocal support from them’.⁸⁶ This quote provides striking support for the assuring power of client commitments that H1 expects. The United States would calculate its involvement ‘pending agreement upon collective defense measures’.⁸⁷ The core mechanism whereby clients need to assure the patron to overcome entrapment concerns, as H1 expects, is nicely brought out by the US ambassador to Britain’s recollection of his discussion with Bevin:

He did not mean that the UK would not stand up and fight, but that although she would fight, she would find it impossible to hold out for a period of two years until we came in. He … gave assurances that although the UK would fight with determination, she was relatively so much weaker now than in 1940 that she could not hope for a successful stand against the Soviet hordes for a protracted period while we were considering whether we would participate actively.⁸⁸

By mid-May, the US ambassador to Britain was able to inform Marshall that the Brussels powers had agreed to pool, standardise, and develop military interoperability. If Soviet forces attacked, ‘the five powers are determined to fight as far east in Germany as possible … their preparations are therefore aimed at holding the Russians … in such a way that sufficient time for the American military power to intervene decisively can be assured’.⁸⁹ On 23 June, one day after US–British–Soviet discussions on Berlin broke down, leading Stalin to blockade the Western sectors on 24 June, Marshall advised his embassies in Paris, London, Ottawa, Brussels, and Amsterdam that he was ready to participate in top secret talks that stressed ‘continuous and effective self-help’ on European security.⁹⁰ The role of Soviet forces in Germany and presumably Stalin’s Berlin blockade in bringing this about, especially alarming the US joint chiefs of staff and forcing US and French threat perceptions closer together, is nicely brought out by Bidault’s telegram to the US ambassador:

As the French government sees it, the Western democracies are faced with (1) an eventual threat, which is Germany; (2) an actual threat, which is the Soviet Union: (3) an immediate threat which is Soviet action in Germany.⁹¹

Bidault’s telegram provides striking support for H6 and the Soviet threat in bringing the Europeans together. As Westad pointed out, ‘in case of war the Red Army would be heading straight for Paris. American influence may be a danger to France’s soul, but Soviet power was a danger to its heart.’⁹² Norway, Denmark, Iceland, Portugal, Ireland, Canada, and Italy were considered as potential alliance partners to the US and Brussels treaty members on the grounds of their ‘stepping-stone’ status that made them vital for communication and defence.⁹³ Soviet occupation of any of them and their offshore territories would have severely increased the threat of occupation of the five Brussels powers. That the United States wanted these countries that had not committed to each other’s security in the NATO alliance is contrary to the theories developed here.

A 20 March 1949 State Department document could state that West European clients

have already taken action to further the security of the North Atlantic area … The military budgets already carried by many of these countries, despite the tremendous load of economic recovery expenditures which they are undertaking, are an added expression of their intention of helping themselves and of not relying solely or even principally on United States assistance to maintain their own security and that of the North Atlantic area.⁹⁴

As Baylis noted, the claim that Bevin ‘charted a deliberate, consistent and foresighted course from the end of the war to the signing of the North Atlantic Pact is not borne out by the documents now available’.⁹⁵ European initiatives towards collective security in the face of the credible Soviet threat assured the United States of these states’ resolve to address the Soviet threat and incentivised the Truman administration to make similar commitments to the group.⁹⁶ European collective security was first and foremost a local initiative driven by the credible Soviet threat of military occupation.⁹⁷ H5 finds support in Truman telling congress that ‘the determination of the free countries of Europe to protect themselves will be matched by an equal determination on our part to help them protect themselves’.⁹⁸

Asia: No credible Soviet threat, no collective security

This section shows that the absence of a credible Soviet threat of military occupation in Asia left US clients there unconcerned about and uncommitted to each other’s defence. It tests H3 and H4 through showing that when the United States attempted to sell collective security to and then impose it on its clients in Asia in 1951 and 1954, these efforts failed and forced the Truman and Eisenhower administrations to have to settle for bilateral alliances. The analysis also shows that geography was not determinative of alliance design because Stalin prepared for a substantial mobilisation of Japan but ultimately called it off. The section further shows that while there is much evidence that racism influenced US policy towards Asia, this does not explain why the Truman administration wanted a multilateral alliance with its clients there.

If geography is sufficient to explain alliance design in Europe and Asia, we should expect no military mobilisations, low perceptions of threat, and no alliances in maritime Asia.⁹⁹ But by the end of the Second World War, a partial Soviet occupation of Japan was in the works. At the 26–7 June combined conference of the Soviet Politburo, government, and military, many wanted the total occupation of Japan’s northern island Hokkaido.¹⁰⁰ By August, Stalin had authorised planning for its partial occupation, and on 16 August wrote to President Truman that he not only wanted the Kurile Islands but the ‘Northern part of Hokkaido’.¹⁰¹ For the Russian dictator, his troops had to have ‘an occupation right in some part of the Japanese proper territory’.¹⁰² The commander of Soviet Pacific Forces in the Far East ordered his general commander on 18 August to occupy Hokkaido.¹⁰³ The plan involved the transport of three rifle divisions, with support from two regiments of twin-engine bombers and four submarines, six destroyers, six torpedo boats, four frigates, four trawlers, and four large hunters. Such maritime power would have been more threatening to US clients in the region had Stalin been able to establish a foothold in Japan. Yet after Truman declined Stalin’s desire for occupation rights in Hokkaido, the Soviet leader called off his occupation plans.¹⁰⁴

The key (missing) data point regarding the prospects for multilateral alliances in Asia in the early Cold War was the absence of any credible Soviet threat of military occupation to focus the minds of US clients in that region. The Soviet mobilisation for but ultimate non-occupation of Japan left US clients in Asia less wary of Soviet military occupation and, as a result, less committed to collective security. The US ambassador to London reported in 1949 that the ‘only Asia-Pacific nations with ‘interests and ideals’ in common with Washington, capable of ‘effective self-help and mutual aid’ were the British Dominions (Australia and New Zealand) and the Philippines’.¹⁰⁵ The prospects of an alliance between Canberra, Wellington, and Manila were, however, slim. Taiwan and the Philippines developed a proposal for Pacific Union in the summer of 1949 but found little regional interest.¹⁰⁶ ‘Diversity of interests’ ensured that the regional response was ‘only lukewarm’, with many opposed to a military alliance.¹⁰⁷ The Baguio summit – held weeks before the Korean War began and reflecting the greatest commitment by Washington’s regional partners to commit to address security challenges – was marred by divergent viewpoints regarding what to prioritise and what strategies to pursue. According to the US ambassador in Australia, Canberra’s disinterest in the Baguio summit, which it had initially declined to attend, stemmed from the fact that its ‘purposes seemed unclear to Filipinos themselves’.¹⁰⁸ By the summer of 1950, Acheson received a report from his ambassador in Pakistan that ‘Baguio conference revealed that time is not yet for meaningful Pan Asiatic consideration of either Asian or non-Asian problems by Asian countries, with exception inoffensive (to west or east) economic and commercial development projects’.¹⁰⁹ Philippine president Quirino himself was apparently most interested in economic and cultural cooperation, crushing the Huk insurgency, and combating arms smuggling and other domestic unrest that threatened his hold on power.¹¹⁰ The Soviet threat did not loom large.

If any Australians were to perceive threat from Moscow, it would be the military chiefs of staff, defence secretary, and their associates, who monitor the balance of military capabilities. These analyses, however, tended to only see the Soviets as threatening if Stalin acquired greater control over Japan. The defence chiefs were especially concerned with Moscow’s influence in Japan and China. The 1950 Defence Committee report, written before the outbreak of the Korean War, noted that ‘the present military threat to the Pacific and the Far East is from Russia and China’ and that future Japanese alignment with the communist bloc was presaged to be ‘a threat to the Pacific area’.¹¹¹ The report concluded that Moscow’s limited ability to neutralise US influence in Japan left a ‘very limited’ threat to the Pacific. Had Stalin managed to muscle his way into the Japanese occupation, Australia’s military chiefs would likely have perceived greater threat from Moscow. The Korean War, moreover, did not change threat perceptions in Canberra. With Soviet forces out of Japan and, it seemed at worst, providing limited air support in the Korean theatre, Canberra did not perceive great Soviet threat to Australia from the war and fought in the war with the principal goal of leveraging a US defensive alliance commitment.¹¹² Australians at the time widely perceived Japan as a ‘known evil’ and the Soviet Union as ‘an uncertain and possible lesser one’.¹¹³ Dulles duly informed the joint chiefs after returning from his trip to Asia of disinterest in Canberra, Wellington, and Manila in an alliance with each other.¹¹⁴

H1 finds support insofar as when the Truman administration concluded after the Korean War broke out that it needed to firm up its defence relationships with clients in Asia, there was a consensus that these states’ commitment to collective defence would assure Washington regarding entrapment concerns.¹¹⁵ Many would have agreed with this before the Korean War. The National Security Council wrote to President Truman at the end of 1949 that any association in Asia ‘must be the result of a genuine desire on the part of the participating nations to cooperate for mutual benefit’ and ‘operate on the basis of mutual aid and self-help in all fields’, and that ‘the United States must not take such an active part in the early stages of the formation of such an association’.¹¹⁶ For Dean Acheson in early 1950, ‘indigenous motivation’ was essential.¹¹⁷ As Truman informed Philippine president Quirino in February 1950, his administration would be ‘most sympathetic toward an Asian Union which was inclusive and which sprang from Asian initiative’.¹¹⁸ As the US ambassador to Thailand noted, given ‘slight Asian interest US cannot push or sponsor association although such association would be in our interest’.¹¹⁹

H1 also finds support in all US reference to alliances in Asia preferring one multilateral alliance. US memos in the weeks before the February 1951 Canberra meetings consistently spoke of one multilateral alliance including Australia, New Zealand, the Phillippines, and Japan.¹²⁰ In January 1951, Dulles, now Truman’s special representative, was commissioned to encourage a ‘mutual assistance arrangement’ between Australia, New Zealand, the Philippines, Japan, and possibly Indonesia.¹²¹ By May 1951, the National Security Council reported that key to US strategy in Asia was ‘the firm establishment and effective application of the principle of collective security … on the basis of self-help and mutual aid’.¹²² NSC 48 would call for the maintenance of ‘the security of the off-shore defense line: Japan–Ryukus–Philippines–Australia and New Zealand’ and for Washington to ‘proceed urgently to conclude a peace settlement with Japan’ that would, as H1 expects, lock in one multilateral alliance with these clients over the longer run.¹²³ Amid client disinterest in collective security, the consistent US preference was, as H1 expects, client commitments to collective security that would assure Washington of shared resolve.¹²⁴ Hemmer and Katzenstein’s argument that US policy towards Asia was marred by racism finds support in Acheson’s report that President Quirino’s delegation were a ‘puzzled and bewildered group of men’.¹²⁵ But this framework cannot explain why Acheson wanted the Philippines in a multilateral alliance. That the US goal was for this alliance to involve Manila, Tokyo, Canberra, and Wellington suggests that there were significant limits on the impact that racism imposed on US alliance strategy. Problematically for the racism thesis, US thinking regarding alliances with the Philippines and Japan evolved over time in very similar ways to that of Australia and New Zealand and the Europeans. The desire for local collective defence is also not what models of alliances as bilateral controls would expect.

H1 finds less support in the Americans seeking a multilateral alliance in Asia when its clients there harboured no such interest. But the US insistence on such an alliance in the face of local disinterest provides further evidence for the assuring power of client commitments to collective security. It also provides a test of H3 and H4 and shows, consistently with these hypotheses, that when a US preference for multilateralism does not match client preferences, multilateralism will fail, and bilateralism will result. Mabon thus noted that US bilateral alliances in the Pacific were motivated by a guiding principle of multilateralism and in this sense were ‘multilateral designs which were unilaterally conceived’.¹²⁶ As H5 expects, the US commitment to its Asian allies to only consult matched client commitments, presumably to extract a greater client collective security commitment. Dulles and his colleagues hoped that their alliances in Asia would foster greater client commitment to others’ defence. He viewed them as potential for ‘more vigorous measures of collective self-defense’.¹²⁷ Contrary to Cha’s power-play thesis, there was no place in US thinking to exert greater control over these status-quo-oriented clients.¹²⁸

H1 also finds less support in the United States not waiting for its allies and partners in Southeast Asia to commit to each other’s defence in 1954 but rather imposing multilateralism on them. The SEATO alliance begs the question of why the Americans put the cart before the horse. The SEATO alliance confirms H3; while Dulles warned that these states’ approach to SEATO would influence the US approach to their prior alliances, insufficient perceptions of communist threat ultimately caused SEATO to flounder and collapse. SEATO begs another question: why did US allies and clients in Asia sign up to SEATO given their very mixed assessments of Chinese and Soviet power? Their commitment to this alliance lay in the fact that most Asian partners – Australia, New Zealand, and the Philippines, but not Thailand, the only SEATO member from Southeast Asia¹²⁹ – were already US allies that viewed SEATO as a means to strengthen their extant alliances. The US had signalled that their commitments to SEATO would have implications for their extant alliances. The joint chiefs of staff, for example, argued in April 1954 that ‘bilateral and multilateral treaties among the countries of the area should be fostered’ to ‘make possible eventual contributions to a collective effort’ involving ‘improved cooperation, coordination of plans, and eventually to a comprehensive and cohesive system of security in the Far East area’.¹³⁰ Canberra, Wellington, and Manila worried that if they were perceived in Washington to commit too little to SEATO, the United States might commit too little to their own alliance.¹³¹

Conclusion

The above theory and empirical analysis have shown that a common strategic logic explains multilateralism in Europe and bilateralism in Asia. As a State Department memo noted, ‘United States purposes in Asia no different from what they are in Europe … (1) standing firm in face threatened aggression, (2) collective security and (3) eschewing aggression’.¹³² That ‘other Western countries with capacity and resources to make contribution to collective defense do not have same Pacific Ocean stakes as United States’ ultimately prevented multilateralism in Asia.¹³³ As President Eisenhower admitted, US policy was most effective when it ‘support(ed) strength where strength exists’.¹³⁴ Patrons contemplating an alliance with their clients face a fundamental entrapment problem. Client commitments to collective security provide assurance of shared resolve that addresses this concern. The analysis has shown that a credible threat of military occupation incentivises clients to commit to each other’s security, and that this is not reducible to geography. If patrons can wait for their clients to commit to each other, they can be part of a multilateral defensive alliance. The analysis has also shown that if they attempt to impose collective security on their clients when an adversary has not already incentivised them to do so, multilateralism will collapse, and the patron will have to settle for bilateral alliances.

H1 found strong support in the consistent US preference for client commitment to collective security in Europe and Asia. H2 also found support in NATO lasting much longer than SEATO. The US encouragement of multilateralism in Asia in 1951 and then the imposition of it on its clients and allies there in 1954 is problematic for H1 but provides strong support for H3 and H4. The strongest forms of collective defence are not imposed from above but indigenously developed, and imposed multilateralism will fail and force a patron to settle for bilateralism. H5 found support in the United States offering the same commitment to its clients that they offered to each other. H6 found support in the credible threat of Soviet military occupation in Europe causing collective security amongst US clients in Europe but not Asia. H7 also found some support in US clients perceiving greater threat in land-based Europe than maritime Asia, but the analysis has shown that geography is not determinative of regional alliance design. NATO may not have exclusively resulted from local European initiative, but the analysis here has shown that the Brussels treaty that preceded the fundamental US security commitment to Western Europe was first and foremost a European initiative. The theory developed here cannot explain the US security commitment to Iceland, Denmark, Norway, Ireland, Canada, Portugal, and Italy, which had not committed to each other’s security. But the theoretical framework here is somewhat consistent with their inclusion insofar as geography meant that their inclusion in what would become the NATO alliance was necessary for the security of the Brussels pact members. H1 cannot explain the US security commitment to its Asian allies before they had committed to each other but explains that multilateralism imposed from above did not work. The US experience in Asia only reinforced the optimal assuring power of client collective security commitments.

The findings have important implications for scholarship and policymaking. Scholarship on the absence of an Asian NATO has grappled with issues of great powers, institutional efficiencies, racism, control, framing, and networks. The analysis here suggests that the answer lies in whether US clients were motivated to commit to each other’s security by the credibility of the Soviet threat of military occupation. The United States in the early Cold War may have been able to assure its clients of its capabilities and resolve, but the design, commitment therein, and longevity of the alliances that emerged hinged on how these clients could assure the United States of their own resolve and capabilities. Such assurance efforts were influenced by Soviet offensive military power and cannot be reduced to geographic determinism; that Asia is a maritime theatre alone tells us nothing about the alliances that would emerge there. A common strategic logic explains variation in alliance design in Europe and Asia and suggests that increasing Russian and Chinese threats of military occupation of US allies and clients will create and firm up multilateral alliances. Russia’s invasion of Ukraine caused Finland and Sweden to join NATO. Chinese aggression in the Indo-Pacific may well make the QUAD increasingly come to resemble a defensive alliance. Under extreme circumstances, US European and Indo-Pacific alliances may become further interdependent. Michael Green pointed out that such ‘networking of trans-Atlantic and Pacific alliances is also a long-standing objective of U.S. strategy’.¹³⁵ Credible threats of military occupation need to be weighed up against the alliances that they likely help create. Stalin substantially complicated US grand strategy for Asia by calling off his planned occupation of northern Hokkaido. One important question for future research is whether the effects found here are restricted to democratic great powers and their clients, which, as Daina Chiba, Jesse Johnson, and Brett Ashley Leeds pointed out, make more careful alliance commitments.¹³⁶ Further analysis could incorporate variables such as ideology and political leaders to further explain variation in alliance design.¹³⁷

The analysis here also has implications for other alliances in multipolarity, bipolarity, and unipolarity. If current G5 Sahel members Mauritania, Niger, Mali, Chad, and Burkina Faso had faced a greater insurgent threat, they would presumably have been more likely to commit to each other’s security. This might have made Paris offer a greater military commitment to this arrangement. Furthermore, the analysis here ultimately suggests that long-lasting alliances in the Indo-Pacific today are more likely if the United States lets its partners and allies take the lead in forming alliances to respond to common challenges. Tracking which recent Chinese policies in the South China Sea, Taiwan Straits, and other disputed areas have been preceded and followed by what allied policies would suggest the conditions when China’s usually calculated coercive plays might cause something resembling a balancing collective security system.¹³⁸ The framework developed here could help identify the conditions when, for example, South Korea might abandon its 2017 commitment to China to not join an alliance with the United States and Japan.¹³⁹ The framework here could also help predict the conditions when Japan and Australia – both members of the QUAD – would make security commitments to each other more resembling a defensive alliance that the United States could then join to constitute an Asian NATO.¹⁴⁰ The analysis here could also shed light on the conditions – on the India-China disputed borders and elsewhere – when India might become more attracted to such a grouping. Similar analysis could suggest the conditions when Chinese behaviour in the South China Sea becomes sufficiently threatening to Australia that Canberra might be more attracted to the prospect of an alliance with the Philippines than it was in the early Cold War. This analysis could also suggest what military moves by Russia caused the Scandinavian countries to pool their advanced fighter jets into a single fleet and what further Russian challenges might lead to further European military integration and/or greater US and allied support of Ukraine and other regional (Baltic state?) partners.¹⁴¹ Further addressing the historical and contemporary relationship between patrons, clients, adversaries, credible military threats, and collective security commitments promises to shed new light on longstanding theoretical and policy debates regarding intensifying and moderating interstate geopolitical competition. Better understanding these mechanisms promises a richer understanding of variation in alliance design.