Network Connections and the Emergence of the Hub-and-Spokes Alliance System in East Asia

Why did the hub-and-spokes alliance system emerge in East Asia? Why did a multilateral alliance system not form there during the Cold War, while the North Atlantic Treaty Organization (NATO) was created in Europe? During the early 1950s, the United States successively concluded bilateral alliances in East Asia; it did so with Japan in September 1951, with the Republic of Korea (hereafter ROK or South Korea) in October 1953, and with the Republic of China (hereafter Taiwan) in December 1954. Ever since John Foster Dulles, the U.S. secretary of state during the Dwight Eisenhower administration, referred to these allies as “spokes on a wheel,” the term hub-and-spokes has become a popular metaphor to describe the U.S. alliance system in East Asia.¹

Scholars of international relations theory have been trying to explain the absence of an alliance system in Asia similar to NATO.² This article joins the efforts to address this question, focusing on three U.S. bilateral alliances that constituted the core of the hub-and-spokes system: the U.S.-Japan alliance, the U.S.–South Korea alliance, and the U.S.-Taiwan alliance. Why did these alliances not evolve into a multilateral alliance despite the existence of common communist threats from the Soviet Union, the People’s Republic of China (hereafter China), and North Korea?

Realists and constructivists offer explanations for the emergence of East Asia’s hub-and-spokes alliance system. The most prominent explanations in both schools emphasize that the system emerged because it was what the dominant actor—the United States—wanted.³ It is true that the hub-and-spokes system served U.S. interests better than a multilateral alliance would have, given that the latter could have enabled relatively small U.S. allies to constrain U.S. behavior more effectively than the former.⁴ These arguments are fundamentally flawed, however, because, as discussed below, the historical record reveals that the United States desired and sought a multilateral alliance in East Asia until the early 1960s. That is, East Asia’s hub-and-spokes alliance system persisted despite the U.S. preference for a multilateral alliance. The other arguments, which emphasize the role of the U.S. allies, offer only partially accurate explanations as well.

This study shows that social exchange theory, rather than realism or constructivism, is the most promising approach to explain the puzzle of why a multilateral alliance did not emerge in East Asia.⁵ More precisely, I argue that the interaction dynamics through which East Asia’s hub-and-spokes alliance system emerged can be best captured by the social exchange network approach—a specific type of social network analysis (SNA) based on social exchange theory.⁶ This approach emphasizes the role of both system-level and agent-level factors in shaping a network structure, and posits that the form of a social network is shaped by the exchange patterns that develop among actors.⁷ I derive from this approach a theoretical model that explains how a specific form of network may emerge among potential allies, and apply the model to explain the emergence of East Asia’s hub-and-spokes alliance system.

The result of my analysis demonstrates that the preferences and behavior of U.S. allies in the region proved at least as consequential as those of the United States in shaping the hub-and-spokes alliance system, which emerged as an unintended consequence of interactions among them. It was the U.S. allies that actively, and aggressively when necessary, sought strong bilateral security ties with the United States because such ties could provide far more security than security ties formed among themselves. As a result, the degree to which the three U.S. allies felt the need to develop security ties among themselves—the very element that would have transformed the hub-and-spokes system into a multilateral alliance—was inversely influenced by the strength of their respective ties with the United States. Such a linkage constitutes what social exchange theorists call “negative connections” between the United States’ security ties with its allies and the security ties among its allies; the stronger the U.S. security ties with its ally developed and the more the ally’s security needs were satisfied, the weaker the incentives the ally felt for strengthening security ties with the other U.S. allies.⁸

This study makes three contributions to the study of alliance politics and East Asian security. First, it adds to the hotly contested debate on the origin of East Asia’s hub-and-spokes system. It highlights the weaknesses of both realist and constructivist explanations, and provides a more historically accurate explanation.⁹ Second, it demonstrates the utility of SNA—the social exchange network approach, in particular—to alliance politics. Regarding an alliance as an exchange of security goods and focusing on how alliance ties are connected, this study demonstrates that the social exchange network approach can highlight aspects of alliance politics that are overlooked by existing alliance theories. Third, from a practical policy perspective, understanding the origin of East Asia’s hub-and-spokes system enables scholars and decisionmakers to identify how and why the U.S. alliance system in the region is changing and to devise appropriate policy responses. Properly understanding the dynamics of alliance politics in East Asia is all the more important now because how to conduct U.S. alliance policy will be an urgent task for the next administration, regardless of the outcome of the November 2020 presidential election. As elsewhere, the constant attacks against alliances by President Donald Trump have made the future of U.S. alliances in East Asia increasingly uncertain.¹⁰

In the following section, I examine the existing explanations for the absence of a multilateral alliance in East Asia and discuss their limitations. I then present a theoretical model based on the social exchange network approach and derive from it hypotheses on the preferences and behavior of the United States and its allies. Next, I test these hypotheses by analyzing the policies of the United States and its allies during the early Cold War period. In the subsequent section, I show how their interactions led to the emergence of East Asia’s hub-and-spokes alliance system. I conclude by discussing this study’s theoretical and policy implications.

There are generally two categories of explanations for East Asia’s hub-and-spokes alliance system. One is based on a constructivist approach, which emphasizes the role of ideational and normative factors.¹¹ Christopher Hemmer and Peter Katzenstein argue that the lack of a collective identity between the United States and its Asian allies has been an important cause of the absence of a multilateral alliance.¹² Using social identity theory, Hemmer and Katzenstein claim that, while U.S. policymakers felt comfortable enough to develop a multilateral alliance with their European allies given their sense of shared collective identity, they were reluctant to do so with Asian allies because they regarded Asians as different or even inferior. Criticizing Hemmer and Katzenstein’s view as U.S.-centric, Amitav Acharya argues that a region-specific, Asian norm that prioritizes the nonintervention aspect of state sovereignty has made Asian states reluctant to accept institutionalized collective security arrangements and, thus, has made a NATO-type multilateral alliance in Asia impossible.¹³ He claims that Asian states were averse to great power intervention as a result of the experience of Western colonialism and that they viewed a collective security arrangement as a new form of control by the Western powers. This hypersensitivity to any sign of infringements by the West on their newly attained sovereignty doomed the Southeast Asian Treaty Organization (SEATO), a multilateral alliance that never functioned properly from its inception in 1954.¹⁴ The other constructivist explanation emphasizes the role of historical memory.¹⁵ John Duffield, for example, argues that the historical memories of Japan’s atrocities before and during World War II made other states in the Asia-Pacific reluctant to endorse the creation of a multilateral alliance that would include Japan.¹⁶

In contrast to constructivists, realists offer explanations that focus on the material capabilities and intentions of the United States. The most influential realist explanation posits that the United States preferred the hub-and-spokes system because it, rather than a multilateral alliance, maximized U.S. influence over its allies.¹⁷ This argument is articulated theoretically by Victor Cha, who puts forward the “powerplay” thesis.¹⁸ In line with the argument that great powers use alliances to constrain small powers,¹⁹ this thesis posits that the United States created the bilateral alliances to restrain South Korea and Taiwan from initiating actions that might entrap the United States in an undesired war with the communist bloc. Cha is careful to point out that the United States did seek to create a multilateral alliance in 1950–51—the Pacific Ocean pact, which would have included Australia, Japan, New Zealand, and the Philippines—but portrays the United States as primarily interested in “winning Japan” as an ally and claims that “the Pacific Ocean pact was window dressing.”²⁰

While various constructivist and realist perspectives offer plausible explanations, they have serious limitations. First, Hemmer and Katzenstein’s argument does not stand up to empirical tests, because U.S. archival records clearly show that the United States did pursue the creation of a multilateral alliance even after the hub-and-spokes system had emerged. As elaborated below, the Eisenhower administration seriously considered creating a multilateral system, dubbed the Western Pacific Collective Security initiative, during the 1950s.²¹ The administration’s National Security Council (NSC) documents frequently referred to the goal of realizing the initiative, indicating the significance of this aspiration. This fact suggests that the lack of common identity between the United States and its allies was insufficient to stop Washington from seriously seeking to create a multilateral alliance in Asia.

Acharya’s argument also contradicts historical evidence, which suggests that some East Asian states were, and are still, willing to accept serious limitations on their sovereign rights in return for U.S. security guarantees. South Korea agreed to allow the United States to retain command authority over its military after the cease-fire that halted the Korean War, and Japan remains willing to host a number of U.S. bases. As for the historical memory explanation, although it explains why some states, such as Australia and South Korea, opposed including Japan in a multilateral alliance, it does not explain why others supported Japan’s inclusion despite their experiences of Japan’s imperialism before and during World War II. Arguably, Nationalist China suffered Japan’s atrocities as severely as any because it had been fighting a war of survival on its own soil since the 1930s. The United States also suffered severely from the war with Japan. Nonetheless, they remained highly tolerant toward Japan and desired its inclusion in an Asian multilateral alliance.

The realist argument proves as empirically problematic as the constructivist explanations. First, Cha’s powerplay argument, just as Hemmer and Katzenstein’s, overlooks the Eisenhower administration’s continued efforts to create a multilateral alliance even after the three bilateral alliances were established. The powerplay argument underappreciates this aspect of U.S. policy because it pays attention only to the U.S. archival documents before the completion of the hub-and-spokes system in 1953, but not to those after that. Second, the powerplay argument downplays the significant role that U.S. allies played in creating the hub-and-spokes system in East Asia. David Mabon shows that Australia’s and Britain’s objections were serious obstacles to the Pacific Ocean pact proposal in 1950–51.²² Hideki Kan and Tatsuya Nishida emphasized that the failure of the proposal resulted less from the U.S. rational calculation of its interests than from Japan’s reluctance.²³

Instead of constructivist or realist explanations, I propose a social exchange network approach to explain the emergence of East Asia’s hub-and-spokes alliance system. Below, I briefly discuss the characteristics of SNA and introduce a specific type of network analysis, the social exchange network approach. Then, I derive from the approach the hypotheses on the preferences and behavior of states.

Over the last decade, SNA has become a popular tool in the field of international relations.²⁴ A network is a set of interconnected actors or nodes, which can be individuals, firms, states, or any other actor. They are linked through various types of relations or ties, such as cultural links, international trade flows, or alliance ties. SNA regards the shape of networked relations among actors as a social structure, analyzing how actors’ behavior is influenced by the structure and, in turn, how their behavior may shape the structure.

Although an alliance system can be easily conceived of as a network, the application of SNA to alliance politics began only recently. Most of the relevant works focus on testing such concepts as balance, the idea that if state X has alliance ties with state Y and state Z, Y and Z are likely to become allies with each other.²⁵ Other international relations scholars utilize SNA’s insights to analyze power relations in hub-and-spokes networks. For example, Daniel Nexon shows how a hub-and-spokes network allowed its hub state, Austria, to maintain the Hapsburg Empire consisting of its subordinate states (spokes), and how the development of ties among the spokes enabled them to revolt against the hub.²⁶ Nexon and his coauthors also point out that the existing U.S. alliance system resembles a typical empire system, as they both exhibit the characteristics of a hub-and-spokes network between the hub state and its allies.²⁷ The existing studies remain silent, however, on how such networks emerge in the first place.

The social exchange network approach may fill this gap.²⁸ Social exchange theory posits that actors are not self-sufficient, and that relations among them develop as they exchange the resources they need from one another.²⁹ In general, the theory assumes that any exchange entails an opportunity cost, and that actors rationally seek to optimize the net benefits through exchanges.³⁰ This assumption suggests that an actor must carefully choose an exchange partner who can offer the largest net benefits—the sum of benefits gained from the partner minus the resources that it may offer in return—while forgoing less rewarding exchange opportunities with other partners. The theory also assumes that the law of decreasing marginal utility applies to any exchangeable resources, indicating that actors are highly motivated to obtain the resources that they lack, whereas they are less motivated to do so when sufficient in such resources. Applying these building-block notions to network analysis, the social exchange network approach points to two factors that shape an emergent network among actors. One is the distribution of reward resources among actors. Because actors seek to maximize their net gains through exchanges, this factor tends to induce denser exchanges among resource-rich actors and to diminish those among others.³¹ The other factor is an actor’s attempt to shape the network structure surrounding it. Not all actors may be satisfied with the network structure likely to emerge as a result of the distribution of reward resources, and they may take actions, using carrots or sticks, to shape the network structure in their own favor.³² Richard Emerson and his students extend these logics to analyze relations among three or more actors and demonstrate that the approach can explain the development of a complex network of relations.³³

As Glenn Snyder explicitly states that “an alliance is an exchange,” relations among allies can be conceptualized as exchanges of various security-related resources.³⁴ Such resources include defense commitments, military aid, arms transfers, basing rights, access to geostrategically significant locations, and control over allies, all of which constitute a state’s capability to provide security for other states.³⁵ Thus, the social exchange network approach is highly applicable to explaining the formation of a network among allies.

Assume that there exist four potential allies—A, B, C, and D—that face common external threats, although to different degrees. Among them, A possesses by far the dominant military and economic resources that enable it to provide security for the others. Assume also that B possesses the second highest amount of resources, whereas C and D are endowed with fewer resources than B. These assumptions are adopted to make the model similar to the situation in East Asia. All four states try to meet their security needs by exchanging various kinds of security-related resources. The more security exchanges they conduct, the stronger are the security ties that develop among them.

Given the distribution of capabilities to provide security among the four, B, C, and D prioritize developing security ties with A rather than among themselves, and they all desire to establish security ties with A more strongly than A would desire to do so with any one of the three (spoke’s preference hypothesis). This is because B, C, and D could benefit much more from security ties with A than A could with any of the three. On the other hand, A prioritizes developing security ties with B, which can provide the largest security benefits among the three (hub’s preference hypothesis). As a result, it is highly likely that the strongest bilateral security ties are likely to emerge between A and B. Once security ties emerge between A and B, the strength of A’s security ties with C and D is inversely related to, or negatively connected with, to borrow the term from social exchange theory,³⁶ the strength of its security ties with B (hub-spoke negative connection hypothesis). This is because A becomes less willing to strengthen security ties with C and D when it obtains sufficient security benefits from B, but becomes more willing to do so when it cannot obtain sufficient security benefits from B.³⁷

Even when A remains unwilling to strengthen security ties with C and D, the latter two will not simply give up. Instead, obtaining A’s security commitments is so important for C and D that they are likely to initiate actions, or what international relations scholars call “binding strategies,” to do so.³⁸ Although a state usually uses “carrots” (reward binding) to obtain stronger defense commitments from its potential ally, C and D may have no choice but to use “sticks” (coercive binding) despite the risk of infuriating and further alienating A (coercive binding hypothesis).³⁹ According to social exchange theorists, an actor highly dependent upon another but lacking the resources that may induce the latter’s cooperation—just like C and D in this situation—becomes willing to use coercive means to obtain the latter’s cooperation because the success of coercion can greatly improve the existing situation, while its failure may worsen the prevailing situation only marginally.⁴⁰

As for security ties among B, C, and D, their willingness to strengthen such ties among themselves is negatively connected with the strength of their respective security ties with A; the more security commitments B, C, or D obtains from A, the less willing each of them becomes to strengthen security ties with the other two, and vice versa (inter-spoke negative connection hypothesis). A simple cost-benefit calculus explains this logic. When each of the three obtains sufficient security commitments from A, the value of the additional security that each of them obtains from the other two is marginal, while the cost of providing security for the other two tends to be substantial for each of the three, given its limited capabilities to provide security. When B, C, or D cannot obtain sufficient security commitments from A, in contrast, its incentives to strengthen security ties with the other two increase. This is partly because its insecurity makes the additional security benefits that it obtains from the other two highly valuable, and partly because strengthening ties with the other two enhances its relative bargaining position vis-à-vis A in persuading the latter to provide more security for itself.⁴¹ That is, B, C, and D will become willing to develop a collective security mechanism when their security is not assured by A.⁴²

It is important to note that this theoretical model does not predict that A will oppose the strengthening of security ties among B, C, and D. A’s attitude toward security ties among the other three is determined by the likelihood that such ties will generate security benefits for A. That is, A supports stronger ties among B, C, and D if such ties create positive security externalities for A, whereas A opposes such ties if they reduce the overall security benefit for itself.

Below, I analyze the process through which East Asia’s hub-and-spokes alliance system emerged and stabilized during the 1950s and the early 1960s to show how the observations obtained from the case studies fit the hypotheses stated above.⁴³ By doing so, this study also serves as a plausibility probe for the social exchange network approach as a potential tool to explain alliance politics.⁴⁴ In conducting case studies, I use both process tracing and congruence methods. Process tracing is an ideal method to examine the underlying causes of state behavior, and I make the most of it to reveal the preferences and intentions of the four states.⁴⁵ The method may not generate sufficient evidence, however, because available sources may be limited. In such cases, I use the congruence method and examine the degree of fitness between the values of causes and effects expected by the hypotheses.⁴⁶ More precisely, I first assess whether and why the United States preferred the hub-and-spokes alliance system or a multilateral alliance in East Asia. I also analyze U.S. behavior toward its allies, and whether or how U.S.-Japan bilateral relations influenced Washington’s policies toward South Korea and Taiwan. I then analyze the behavior of the three U.S. allies toward the United States and one another. In doing so, I utilize archival materials collected in the United States, Japan, and Taiwan, as well as Korean archival materials available online, while making use of recent secondary sources that employ non-U.S. archival materials.

Both major realist and constructivist explanations posit that the United States preferred a hub-and-spokes alliance system to a multilateral system. This view is fundamentally flawed, as U.S. archival documents show that Washington preferred a multilateral alliance to the hub-and-spokes system in East Asia even after the latter emerged.

When the Cold War required the United States to revise its East Asian security strategy, U.S. policymakers’ main focus was Japan. By the summer of 1947, George Kennan, the director of the Department of State Policy Planning Staff, keenly recognized Japan’s geostrategic significance and opposed what he viewed as an outdated policy of Japan’s complete disarmament. He argued that Japan should be defended by the United States to deny it to the Soviet bloc.⁴⁷ The leaders of the U.S. military shared this opinion and went even further than Kennan’s assessment. According to NSC 49, which was prepared by the U.S. Joint Chiefs of Staff (JCS) in June 1949, Japan was “of high strategic importance to United States security interests in the Far East … because of [its] geographic location” and its “manpower and … industrial potentials.”⁴⁸ The document further stated that “Japan [could] be expected, with planned initial U.S. assistance, at least to protect herself and … to contribute importantly to military operations against the Soviets in Asia.“⁴⁹ Accordingly, the United States contemplated making some form of security arrangement with Japan even before the outbreak of the Korean War.⁵⁰

In contrast, the U.S. government originally recognized only limited strategic value in South Korea and Taiwan. In September 1947, the JCS confirmed its earlier assessment that “the United States ha[d] little strategic interest in maintaining the present troops and bases in Korea” and that “the withdrawal of these forces would not impair the military position of the Far East Command.”⁵¹ Accordingly, the U.S. government pursued the establishment of a neutral and unified Korea through the United Nations (UN), and completed the withdrawal of its troops there by July 1949. The U.S. government also assessed that the strategic value of Taiwan was not sufficient to make U.S. defense commitments there. NSC 48/2 stated that “the strategic importance of Formosa [i.e., Taiwan] does not justify overt military action” and that “the United States should make every effort to strengthen the over-all U.S. position with respect to the Philippines, the Ryukyus, and Japan.”⁵² These assessments were behind the famous speech by Secretary of State Dean Acheson in January 1950, in which he excluded both Korea and Taiwan from the U.S. defense perimeter in the Pacific.⁵³

It is important to note, however, that U.S. policymakers did not regard bilateral security cooperation with Japan as an end in itself. NSC 125/2, approved in August 1952 under Harry Truman’s administration, stated that a “strong, stable, and independent Japan restored to an influential position in Asia could be the most effective ally of the United States in Asia.” This assessment was accepted because “Japan’s military strength,” when developed, “can contribute to the security of the free nations of the Pacific area and of the northern portion of the off-shore island chain,” thereby reducing U.S. security burdens in the region.⁵⁴ Furthermore, NSC 125/2 states that “the United States should encourage and where desirable participate in collective security arrangements in the Pacific area which would include Japan as an important member. Such arrangements would facilitate Japan’s contribution to the security and economy of the free nations of the area.”⁵⁵ These statements reveal the U.S. desire for developing a multilateral security system in East Asia, and that Japan’s significance for the United States lay in its ability to contribute to regional security in East Asia.

The U.S. desire for a multilateral alliance in East Asia became more evident under the Eisenhower administration. NSC 5416, prepared by the JCS in April 1954, lamented that “too little [effort was made] upon the development of the collective military capabilities of the Asiatic non-Communist countries” and expressed that “the emergence of a regional security pact” was required to accomplish U.S. security objectives “without ever-increasing demands upon United States resources.”⁵⁶ The document stipulated that the critical piece of the regional security pact was “the healthy development of the Japanese military,” which would enable Japan to “become capable of providing for her own security and of becoming a contributor to collective security in the Western Pacific.”⁵⁷ This view was accepted by the NSC and reflected in NSC 5429/5, the administration’s comprehensive Asia policy approved in December 1954. It stipulated that the United States should “encourage the conditions necessary to form as soon as possible and then participate in a Western Pacific collective defense arrangement including the Philippines, Japan, the Republic of China, and the Republic of Korea, eventually linked with the Manila Pact and ANZUS.”⁵⁸

What the United States envisioned was a defensive multilateral system that was expected to reduce its security burdens in the region. In accordance with the “New Look” policy, the Eisenhower administration aimed to limit military spending to an acceptable level, and it planned to allow its local allies to assume more costs and defense obligations, the provision of land troops in particular.⁵⁹ A multilateral security framework, in which U.S. allies contributed to each other’s security, was thus considered more desirable than a hub-and-spokes system.⁶⁰ On the other hand, the U.S. government recognized that the negative effects of promoting such an alliance were the heightened risk of entrapment into undesired wars triggered by anti-communist states such as South Korea and Taiwan, and the resulting divergence between those anti-communist states and neutral states.⁶¹ How to manage these potential problems would complicate the U.S. pursuit of a multilateral alliance in East Asia, as discussed below.

To achieve the goal of forming a multilateral alliance, the Eisenhower administration undertook concrete initiatives. One was the Mutual Security Assistance Agreement with Japan signed in March 1954. By concluding the agreement, the United States aimed to encourage Japan to accelerate its military buildup. Another initiative was the U.S. attempt to promote a Seoul-Tokyo rapprochement. The Eisenhower administration regarded the rapprochement between the two countries as the first step for enhancing multilateral security cooperation in East Asia, and actively encouraged them to normalize relations. Its efforts did not bear fruit, however, as U.S. diplomacy met strong resistance from both states. For instance, when President Eisenhower invited South Korean President Syngman Rhee to the White House in July 1954 and asked him to normalize relations with Japan, Rhee adamantly refused, making Eisenhower angry.⁶² The John F. Kennedy administration invigorated the efforts to bridge the gap between Seoul and Tokyo, but it was only in 1965 that the U.S. government finally succeeded in getting its two allies to normalize relations.⁶³

Victor Cha argues that controlling Japan bilaterally was a persistent goal for the United States after World War II.⁶⁴ It was not the United States but Japan, however, that revealed a strong preference for security bilateralism, and it strongly resisted U.S. efforts to multilateralize East Asia’s alliance system. Ironically, what enabled Japan to free-ride on U.S. security efforts was the strong U.S. alliance commitments to Japan and U.S. alignments with South Korea and Taiwan.

Japan first indicated its desire for a U.S.-Japan bilateral alliance as early as September 1947, when Foreign Minister Hitoshi Ashida handed to the U.S. side the so-called Ashida Memorandum, which signaled Tokyo’s willingness to conclude a bilateral security treaty.⁶⁵ While the Japanese government had believed that it would have to accept a restrictive and punitive security arrangement at the beginning of the U.S. occupation, the steadily intensifying Cold War encouraged Japan to explore various options for its post-occupation security. Although Ashida’s message did not solicit a meaningful U.S. response, Japan’s Ministry of Foreign Affairs concluded by December 1949 that some form of security arrangement with the United States would best serve its needs, according to studies based on Japanese archives.⁶⁶

An important impetus for a U.S.-Japan alliance came from Tokyo in May 1950, when Japan’s prime minister, Shigeru Yoshida, instructed his protégé, Finance Minister Hayato Ikeda, to secretly inform the U.S. government that Japan was prepared to conclude a security treaty.⁶⁷ Significantly, Yoshida indicated his willingness to keep U.S. bases in Japan after the end of the U.S. occupation, even offering “to ask the United States to retain its bases in Japan from the Japanese side if the U.S. side find it difficult to make such a request.“⁶⁸ Yoshida knew that the status of U.S. bases in Japan had been the stumbling block that plagued U.S. policymakers. Although a consensus had already emerged within the U.S. government that some kind of security arrangement with an independent Japan was necessary, U.S. policymakers disagreed on how to proceed because they believed that once sovereignty was restored, leaders in Tokyo would oppose the retention of U.S. bases on Japanese territory.⁶⁹ Yoshida’s offer thus served as the ice breaker, enabling the Department of State to persuade the Department of Defense to sign the memorandum supporting Japan’s independence on June 23.⁷⁰

From October 1950 to January 1951, senior diplomats, Yoshida’s security policy advisers, and Yoshida himself held a series of meetings to decide how Japan should present its post-independence security policy to the U.S. delegation scheduled to arrive in late January 1951. During the process, the group’s focus was on how to secure the strongest possible U.S. security commitments to Japan. Early on, Yoshida expressed skepticism of the United Nations and instructed diplomats to create “a framework that would completely satisfy Japan’s security,” meaning a U.S.-Japan bilateral security framework.⁷¹ Yoshida’s advisers frequently expressed the fear that the United States might abruptly withdraw its troops from Japan. Thus, they argued that Japan should “bind the United States [to defending Japan] via a treaty in order to prevent U.S. unilateral disengagement,”⁷² and that “the treaty should be designed to provide as many benefits as possible to the United States.”⁷³ Ministry of Foreign Affairs officials took this advice and wrote up a draft U.S.-Japan treaty.

When the U.S. delegation led by John Foster Dulles, the consultant to the secretary of state on the peace treaty with Japan, came to Tokyo, the Japanese negotiating team presented the aforementioned draft on February 1, 1951. The U.S. side responded favorably, and they decided to pursue a bilateral security arrangement based on this proposal. When a U.S. delegation member “inquired as to whether Japan was willing to contribute to such a regional collective defense pact,” the Japanese side immediately dismissed it, maintaining that Japan “[could] not go beyond what it would do under the U.S.-Japan cooperative accord that had been discussed.”⁷⁴ Reflecting on the entire negotiations for the peace treaty and the security treaty, Kumao Nishimura, Japan’s senior diplomat involved in every aspect of the preparations and negotiations for the peace and security treaties, claimed that the bilateral security treaty was Japan’s initiative.⁷⁵

While the U.S.-Japan security treaty, signed in September 1951, had some problems from Japan’s perspective, the Japanese government was generally content with the level of U.S. security commitments. The treaty guaranteed the United States the right to intervene to quell domestic disturbances within Japan, on the one hand, but did not clearly stipulate a U.S. obligation to defend Japan.⁷⁶ The opposition parties criticized Yoshida for these shortfalls, calling the treaty “unequal.” While the Japanese officials who had negotiated the treaty with the United States recognized these problems, neither they nor Yoshida considered these problems as reflecting the weakness of U.S. commitments. They were fully aware that the U.S. government evaluated Japan’s geopolitical value as significant. Indeed, NSC 125/2, as well as many other documents, states that Japan was so significant for U.S. security that “the United States would fight to prevent hostile forces from gaining control of any part of the Japanese territory,” and Japanese officials were well aware of the U.S. policy.⁷⁷ Thus, for Japanese officials, it was unthinkable that the United States would not respond militarily if Japan, where U.S. forces were stationed, were attacked.⁷⁸ After all, these shortfalls were resolved when the security treaty was replaced in 1960 by a new treaty, which stipulated the U.S. defense obligation toward Japan while omitting the U.S. right to intervene domestically.

In contrast to its willingness to develop bilateral security ties with the United States, Japan consistently opposed joining any regional security system. Prime Minister Yoshida’s well-known reluctance to develop Japan’s military capabilities was one reflection of such attitudes. During the negotiation process of the 1954 U.S.-Japan Mutual Security Assistance Agreement, for instance, the Japanese government repeatedly demanded the elimination of the term “collective defense” from the agreement, thereby irritating the U.S. side, whose primary goal for the agreement was precisely to encourage Japan to develop its military capabilities for regional collective defense.⁷⁹ When U.S. officials inquired about the possibility of Japan’s joining a northeast Asian collective alliance in July 1954, a senior Japanese diplomat stated that “Japan would not be able to participate in a collective defense organization since Article 9 of the Japanese Constitution [was] generally interpreted to prohibit the sending of Japanese forces abroad.”⁸⁰ U.S. Ambassador to Japan John Allison assessed that “there [was] practically no possibility that Japan at present would consider joining any collective security organization.”⁸¹

Japan’s reluctance to contribute to regional security was most vividly revealed in its attempt to revise the U.S.-Japan security treaty in 1955. In July 1955, Foreign Minister Mamoru Shigemitsu handed a draft of a revised U.S. Japan security treaty to Ambassador Allison and proposed to discuss it during his coming visit to Washington.⁸² The draft was remarkable in that it stipulated that Japan would “act to meet the common danger” in the case of an armed attack against areas such as Guam while requiring the eventual U.S. military withdrawal from Japan.⁸³ The Japanese side subsequently clarified that this treaty “would obligate Japan to send its troops overseas” when conditions were met.⁸⁴ The U.S. government welcomed Japan’s initiative as the first sign that Tokyo had become willing to contribute to regional collective defense.⁸⁵ In fact, this was the only occasion during the entire Cold War period that the Japanese government expressed its intention to join its ally in collective military actions abroad.

It soon turned out that the Japanese proposal fell far short of U.S. expectations, however. When the U.S. side asked if the revised treaty would cover areas surrounding the Korean Peninsula or Taiwan, the Japanese side replied negatively. The Japanese side even stated that it would not defend U.S. military vessels or aircraft in these areas, expressing its desire to avoid the risk of entrapment in crisis situations similar to the Taiwan Strait crisis from September 1954 to April 1955. When the U.S. side inquired as to whether Japan would consider joining a multilateral defense pact in the area, the Japanese side expressed its ambivalence.⁸⁶ The U.S. government concluded that Japan’s true intention was to expedite the withdrawal of U.S. forces in Japan rather than to enhance Japan’s contribution to regional security and thus decided not to respond favorably to the proposal.⁸⁷ Japan’s negative attitude toward a multilateral alliance persisted even after the revision of the U.S.-Japan security treaty in 1960. The Ministry of Foreign Affairs’ internal document prepared in April 1960 stated that Japan could contribute to regional security only through the provision of base facilities for the United States, and that “it was impossible for Japan to join a collective security organization such as a northeast Asian treaty organization.”⁸⁸

Japan also took negative positions toward strengthening security ties with South Korea and Taiwan. During the peace settlement negotiations with Taiwan in early 1952, the Japanese negotiators had been explicitly instructed not to agree to any kind of security arrangement.⁸⁹ When the Taiwanese ambassador to Japan, Dong Xian-guang, met Prime Minister Yoshida and requested that Japan participate in joint anti-communist campaigns in October 1953, Yoshida dismissed the request immediately.⁹⁰ Japan’s attitude toward Taiwan did not change even under the leadership of Prime Minister Nobusuke Kishi, arguably the most pro-Taiwanese Japanese prime minister during the Cold War period.⁹¹ During the Second Taiwan Strait Crisis in 1958, for instance, the Kishi administration repeatedly indicated to the U.S. government its wish to avoid the use of U.S. forces deployed in Japan. Foreign Minister Aiichirō Fujiyama even testified in a Diet session that the Japanese government would restrain U.S. military actions toward the offshore islands in the Taiwan Strait to avoid the risk of entrapment.⁹²

Japan was even less sympathetic toward South Korea. The negotiations to normalize relations between the two states faltered soon after they began in October 1951 and would not resume until 1958. Although the United States regarded the normalization of Japanese–South Korean relations as a vital prerequisite for creating a regional multilateral alliance, it was unable to bridge the gap between Tokyo and Seoul.⁹³ Even when the South Korean government became willing to normalize relations with Japan after the change of governments in Seoul in 1960, the Japanese government under Prime Minister Hayato Ikeda remained wary about pursuing a diplomatic breakthrough.⁹⁴

It is noteworthy that Japan took this stance toward Taiwan and South Korea despite its recognition of their geopolitical significance. For instance, in March 1946, the Ministry of Foreign Affairs viewed that an international security arrangement for the Korean Peninsula was an important element for Japan’s post-occupation security policy.⁹⁵ During the process of formulating Japan’s security policy plan that Tokyo would later present to Dulles, Yoshida also expressed his view that the Korean Peninsula was strategically significant.⁹⁶ As for Taiwan, around the mid-1950s, the Japanese government reached a policy consensus that the separation of Taiwan from China was vital for Japan’s security.⁹⁷ One archival document expressing Japan’s position on Taiwan states that “the acquisition of Taiwan by the Chinese Communists must be absolutely prevented to ensure the security of Japan and the Free world.”⁹⁸ Foreign Minister Shigemitsu re-emphasized what was written in this document to Ambassador Allison when they discussed policy toward China in November 1955.⁹⁹ Ironically, it was U.S. security commitments to South Korea and Taiwan that made it unnecessary for Japan to contribute to the security of the two anti-communist states. Shigeru Yoshida explained that despite the absence of a Korean-Japanese rapprochement, Japan remained secure because of the UN forces’ presence in South Korea.¹⁰⁰

Facing Japan’s reluctance to contribute to regional security and fearing that pressuring Japan to do so would strengthen Japan’s rising nationalism and neutralism, the United States lowered its expectations for Japan. NSC 5913/1, adopted in September 1959, downplayed the possibility of creating a West Pacific collective security body and merely called for the development of “wider understandings of common purposes among all” U.S. allies in the region.¹⁰¹ Although neither the Eisenhower nor the Kennedy administration completely gave up the hope of making Japan assume more security responsibilities for regional defense, the U.S. government began to readjust its security strategy for East Asia.¹⁰²

Conventional wisdom posits that the United States signed an alliance treaty with South Korea to constrain the latter from taking risky actions that could trigger a conflict on the Korean Peninsula.¹⁰³ This view, while partially true, misses the important fact that it was South Korea that created the situation in which the United States only reluctantly offered a bilateral alliance with Seoul. At the same time, Seoul’s willingness to accept the U.S. retention of command over the South Korean armed forces contradicts the regional norm argument emphasizing East Asian states’ sensitivity to sovereignty. South Korea’s objection to including Japan in a regional alliance appears to support the historical memory hypothesis. A close examination of the case, however, reveals that other factors, which the social exchange network approach highlights, were also in play.

When the Korean War armistice negotiations came close to a conclusion in the spring of 1953, the South Korean government reinvigorated its efforts to obtain a bilateral security treaty with the United States. On April 3, 1953, South Korean Foreign Minister Y.T. Pyun, while expressing Seoul’s opposition to the armistice, indicated that a security pact “would be the price of ROK cooperation with armistice efforts.”¹⁰⁴ President Syngman Rhee more explicitly requested, in his April 14 letter to President Eisenhower and in the April 30 letter to UN Cmdr. Mark Clark, that a bilateral defense pact should be concluded as a precondition for Seoul’s cooperation with the armistice.¹⁰⁵ Han Pyo-wook, a South Korean diplomat deeply involved in U.S.-South Korean relations, recollected that Rhee had considered the armistice to be inevitable and that he desired that an alliance treaty be concluded before the signing of the armistice agreement.¹⁰⁶

South Korea found the U.S. responses utterly disappointing, as the U.S. government still considered South Korea to be geopolitically insignificant even at the end of the Korean War. NSC 157/1, adopted in July 1953, is revealing in this regard.¹⁰⁷ The document shows that the U.S. government compared the desirability of two alternatives: “Korea divided … with the Republic of Korea tied into the U.S. security system and developed as a military ally,” or “unified, neutralized Korea under a substantially unchanged ROK” with U.S. forces removed from Korea. After elaborating the costs and benefits of both alternatives, the document concluded that the latter was more desirable for the United States than the former. On May 22, Eisenhower rejected South Korea’s request for a bilateral security treaty while offering alternative plans to address Rhee’s concerns.¹⁰⁸

Unable to persuade the United States to form an alliance, Rhee resorted to a risky unilateral action.¹⁰⁹ On June 18, 1953, he ordered the release of 27,000 North Korean prisoners of war (POWs), whose repatriation had been demanded by the communist side in the armistice negotiations.¹¹⁰ Rhee’s action shocked U.S. policymakers because it could derail the armistice negotiations. They also feared that Rhee might even withdraw the South Korean troops from the UN command and take unilateral military actions against North Korea. The U.S. government contemplated two undesirable options: offering something that would satisfy Rhee in return for his compliance with the armistice, or arranging a coup to replace Rhee with a more controllable leader. After exploring the possibility of the second option, the U.S. government decided to choose the first option and proposed to Seoul a security treaty in July 1953.¹¹¹ Through a difficult bargaining process, the bilateral security treaty was signed in October 1953.

Although no direct evidence exists on Rhee’s motive for releasing communist POWs, at least one of his motives was to force the United States to sign an alliance treaty with South Korea. That South Korea repeatedly demanded a U.S.–South Korea alliance as a precondition for its acceptance of an armistice agreement reflects Rhee’s intention. For instance, soon before the release of the POWs, Rhee put forward another urgent request that the United States conclude an alliance with South Korea before the armistice agreement.¹¹² In addition, both South Korean and U.S. officials point out that Rhee’s real goal was a U.S.–South Korean alliance. Han Pyo-wook recollects that Rhee believed that simply following the United States and cooperating with the armistice negotiation would lead only to South Korea’s suicide. He argues that “by releasing the communist POWs, Rhee demonstrated what he could do to the United States because it had not responded positively to his request for concluding a security treaty.”¹¹³ U.S. Ambassador to South Korea Ellis Briggs argued that Rhee’s risk-taking behavior stemmed from his fear that “Korea might be sacrificed to some great power as happened in 1950” without a security pact with the United States.¹¹⁴ Judging from these observations, it is reasonable to conclude that Rhee’s risk-taking behavior was at least partly a coercive binding attempt aimed at the United States.

For South Korea, the security treaty was an incomplete victory. On the one hand, the treaty guaranteed U.S. security commitments to South Korea’s survival. On the other hand, the treaty made clear that U.S. commitments were merely defensive, obligating the United States to defend South Korea only when the latter was attacked. The South Korean government also had to accept the U.S. retention of operational control over its forces as a means of ensuring that South Korea would not initiate conflict with North Korea unilaterally.¹¹⁵ The U.S.–South Korean alliance thus met Seoul’s primary need for survival, but was insufficient for achieving Seoul’s ultimate purpose of unifying Korea, by force if necessary. What Seoul needed, then, was an additional security system that would enable it to achieve its offensive purpose without sacrificing the existing U.S. security commitments.

After the Korean War cease-fire in July 1953, South Korea renewed its pursuit of multilateral security systems. In November 1953, Syngman Rhee visited Chiang Kai-shek in Taipei, and together they proposed the creation of an “anti-communist united front” in East Asia.¹¹⁶ After the meeting, the South Korean government collaborated with the Taiwanese government to develop an Asian anti-communist collective organization and hosted the meeting among themselves and six other Asian states or entities in June 1954, resulting in the creation of the Asian People’s Anti-Communist League (APACL).¹¹⁷ According to South Korea’s secret instructions to its delegate at the June 1954 meeting, its government considered that “all the Free Nations of Asia should join in guaranteeing eternal security for each other” through a collective security arrangement.¹¹⁸

South Korea also sought an alliance among Taiwan, South Vietnam, and itself. When the South Vietnamese President, Ngo Dihn Diem, visited Seoul in September 1957, President Rhee proposed the creation of “a United Free Asia,” which South Korea expected to be a collective security system consisting of anti-communist states willing to take actions against the communist bloc. Rhee argued that if China took aggressive actions, South Vietnam, Taiwan, and South Korea would retaliate, and thus “a United Free Asia” would be able to defeat and destroy the communist bloc. Diem, however, did not express his consent to the idea, stating that the Geneva Peace Accords prohibited South Vietnam from forming any military alliance.¹¹⁹ Rhee sought another multilateral system in the late 1950s. Under his direction, the South Korean Foreign Ministry sent telegrams, on August 29, 1959, to its embassies in Taiwan, the Philippines, and South Vietnam urging its diplomats to discuss with their host governments the creation of the Organization of East Asian Nations. According to the South Korean documents, this organization was supposed to unite anti-communist states in Asia and to promote military, political, and other types of cooperation among them.¹²⁰ As a result of the Rhee administration’s collapse in 1960, however, no such cooperation materialized.

South Korea also sought to enhance military cooperation with like-minded states in East Asia. In October 1953, Foreign Minister Yung-tai Pyun proposed to the Taiwanese ambassador in Seoul, Dong-yuan Wang, the creation of a secret joint military plan in which South Korean forces would advance into northeast China and Taiwanese forces would attack the Chinese coastal area if a renewed conflict occurred in the Korean Peninsula.¹²¹ During his visit to Taipei, Rhee discussed bilateral military collaboration with Chiang Kai-shek and orally agreed with Chiang that South Korean forces would advance into North Korea in the event of a Taiwan counterattack against mainland China.¹²² Rhee also repeatedly expressed his willingness to send South Korean troops to Indochina in 1954, only to be rebuffed by France, which feared that accepting Seoul’s offer would instead lead to greater Chinese interventions.¹²³

South Korea’s initiatives, however, had a fundamental problem detrimental to the creation of a multilateral security system in East Asia: Seoul adamantly refused to include Japan in any such system. President Rhee consistently opposed the U.S. policy of inducing Japan to contribute militarily to regional security. When Secretary of State Dulles broached the possibility of a multilateral security system including Japan during Rhee’s visit to Washington in July 1954, for instance, Rhee immediately opposed the idea, and expressed his frustration about the U.S. policy that, in his view, excessively stressed Japan’s significance in contrast with that of South Korea.¹²⁴ Rhee also refused to include Japan in the APACL even though the United States and Taiwan advocated its inclusion.¹²⁵ Rhee’s position was not without cost; the United States, because it considered Japan’s participation in a multilateral security system to be of vital importance, declined to participate in the APACL.¹²⁶ Because other states also declined to join the APACL or did so half-heartedly, it remained an ineffective institution.

While Rhee’s and South Koreans’ anti-Japanese sentiments resulting from Japan’s colonial rule over Korea surely played a role in Seoul’s refusal to include Japan in regional security systems, other factors from the social exchange network perspective also account for South Korea’s attitude toward Japan. First, as many scholars point out, South Korea did not urgently need to strengthen its security ties with Japan because it had already attained the alliance with the United States.¹²⁷ Indeed, Han Pyo-wook confirmed that Rhee was in no hurry to restore relations with Japan.¹²⁸ Moreover, South Korean officials feared that strengthening security ties with Japan might encourage the United States to reduce its security commitments to South Korea.¹²⁹ This concern was far from misguided because shifting security burdens to Japan was precisely what the Eisenhower administration had in mind. On September 14, 1954, when John Hull, the commander in chief of the UN Command in Korea, brought up with Kim Yong-shik, a South Korean diplomat, the idea of creating a regional security system in which Japan would play the central role, Kim adamantly opposed it, arguing that what was critical for South Korea was a clear security commitment by no state other than the United States.¹³⁰ Even South Korean President Park Chung-hee, who normalized relations with Japan in 1965, worried about U.S. disengagement after normalizing relations with Japan. President Park expressed such a concern to President Kennedy in his letter dated February 12, 1962, and demanded U.S. reassurance that Washington would not let Japan assume more security responsibility for the Korean Peninsula.¹³¹

Second, South Korea sought to create an offensive multilateral security system and regarded the inclusion of Japan as an impediment. Not completely satisfied with the U.S.–South Korean alliance, which was defensive in nature, South Korea desired to use multilateral security systems to achieve an additional goal of confronting the communist bloc.¹³² To do so, South Korea needed to cooperate with like-minded, anti-communist states, such as Taiwan and South Vietnam, but was reluctant to include Japan and other moderate states because doing so would require toning down the offensive purpose of such a body. In fact, Rhee conveyed to President Eisenhower in December 1954 his openness to a security pact comprising the United States, South Korea, and Japan, but indicated that Japan must be strongly committed to anti-communist causes.¹³³

One episode nicely demonstrates that the factors stated above influenced Seoul’s alliance policy. In late October 1953, Taiwan’s ambassador to Japan, Xian-guang Dong, visited Seoul under Chiang Kai-shek’s instructions and handed a draft of a South Korean–Taiwanese defense treaty to President Rhee. Chiang intended to expand this bilateral treaty to later include other noncommunist states in Asia. Interestingly, Rhee declined Taiwan’s proposal.¹³⁴ Having just signed the alliance treaty with the United States, Rhee argued that the proposed treaty would not have practical effects and that the two states should pursue more aggressive aims against the communist bloc. Therefore, he emphasized the need for an offensive alliance between South Korea and Taiwan and promised that South Korea would support a Taiwanese attack on China. However, he also asked Taiwan to act first because his hands were tied by the security treaty with the United States.¹³⁵

Washington’s attitude toward Seoul gradually changed in a more favorable direction once the U.S.–South Korean alliance formed. Particularly because Japan remained reluctant to contribute to regional security, U.S. policymakers began to find more value in South Korea’s potential, its capabilities to commit land troops in particular.¹³⁶ NSC 5514, approved in March 1955, clearly shows that the U.S. expectations for South Korea’s role in Asia had risen; it set a U.S. policy goal “to enable it to make a substantial contribution to free world strength in the Pacific area” and to include South Korea in a Western Pacific collective defense arrangement.¹³⁷ Secretary of State Dulles, one of the early skeptics of concluding an alliance with South Korea, declared in the NSC meeting on May 28, 1954, that “if an additional division were needed in Indochina he would prefer to have a ROK division go there rather than U.S. ground forces,” and JCS Chairman Arthur Radford immediately agreed.¹³⁸ By 1954, South Korean forces had grown to 650,000 troops in twenty divisions, vastly outnumbering those of Japan, which was claiming that 180,000 troops was the maximum it was willing to develop.¹³⁹ Under such circumstances, the U.S. government concluded that “Korea must be looked on as our ‘force’ in the Far East while no significant increase in support of U.S. objectives can be expected.”¹⁴⁰

In many aspects, Taiwan was situated in a position similar to South Korea; both were divided states, aligned with the United States, and ruled by strong-minded, anti-communist leaders. In contrast with South Korea, however, Taiwan displayed a more tolerant attitude toward Japan, contradicting the expectation of the historical memory hypothesis. This contrast reflects an oft-overlooked difference between Taiwan and South Korea in terms of their security environments.

Taiwan reinvigorated its quest for U.S. alliance commitments after the victory of Dwight Eisenhower in the 1952 U.S. presidential election. Before the Korean War erupted, the Truman administration literally abandoned Taiwan; in his January 1950 speech, Secretary of State Acheson declared that the island was outside of the U.S. defense perimeter.¹⁴¹ Although the outbreak of the Korean War prompted the Truman administration to place the Seventh Fleet in the Taiwan Strait, it did so not to defend Taiwan but to neutralize the strait. Chiang Kai-shek therefore regarded the victory of the pro-Taiwan Republican Party as an opportunity to strengthen security ties with the United States and, in December 1952, set two goals vis-à-vis the United States: an alliance treaty and the continuation of U.S. military aid.¹⁴² In March 1953, Taiwan’s ambassador to the United States, Wellington Koo, proposed to Secretary of State Dulles the idea of signing an alliance treaty. Not having received meaningful responses from Washington despite Taiwan’s repeated requests for a U.S.-Taiwan alliance, Chiang himself presented the idea to U.S. Ambassador to Taiwan Karl Rankin on June 28, 1954.¹⁴³

Although the Eisenhower administration reversed Truman’s policy and integrated Taiwan into the U.S. defense perimeter, it remained reluctant to sign an alliance treaty with Taiwan.¹⁴⁴ When Koo broached the idea of a formal alliance in March 1953, Secretary of State Dulles expressed reservations while touting the need for a regional security arrangement for Asia. Citing the expected complications that might arise by signing a U.S.-Taiwan alliance before the Geneva Conference scheduled for the spring of 1954, the Eisenhower administration decided to postpone discussions on the issue.¹⁴⁵ The Department of State remained reluctant to move forward even after Chiang’s request in June 1954, and Dulles decided on September 1 to delay setting a date for the treaty negotiations with Taiwan.¹⁴⁶ This reluctance was shared by the Pentagon, which opposed increasing defense obligations in the Pacific. Although the U.S. military found some utility in using the Taiwanese military to pressure China along its coastal areas, it estimated that the prospective cost of increasing the United States’ security burdens would outweigh the prospective benefits.¹⁴⁷

The crucial trigger for the U.S.-Taiwan alliance came from neither Washington nor Taipei, but from Beijing. In September 1954, China’s People’s Liberation Army started the shelling of the offshore islands held by Taiwan, initiating the First Taiwan Strait Crisis. Chiang took this opportunity to pressure the United States to sign an alliance treaty. When he met Secretary of State Dulles on September 9, he first raised the issue of an alliance and requested its early conclusion.¹⁴⁸ He followed up his message in a meeting with U.S. Ambassador Rankin on September 21 and “gave [an] unusually strong impression of impatience” about U.S. indecision.¹⁴⁹ To prevent China from attacking Taiwan and to obtain Taiwan’s cooperation for U.S. diplomacy to de-escalate the crisis, the Eisenhower administration decided to conclude a security treaty with Taipei.¹⁵⁰ The two sides quickly began the treaty negotiation, and the mutual defense treaty was signed on December 2, 1954, in the midst of the crisis in the Taiwan Strait.

Although Chiang expressed his sense of “marvelous achievement” in his diary when the alliance treaty was concluded, the Taiwanese side soon became frustrated with the terms, finding them too restrictive and insufficient.¹⁵¹ To conclude the U.S.-Taiwan security treaty, Taiwan had to promise not to conduct large-scale military actions against the mainland without prior consultation with Washington. Furthermore, the treaty left the U.S. obligation to defend the so-called offshore islands ambiguous even though they were under Taiwan’s effective control.¹⁵² The latter point posed a serious problem for Taipei in particular; the loss of the islands would undermine the credibility of Taiwan’s claim that it would ultimately return to the Chinese mainland, which would potentially be detrimental to the Chinese Nationalists’ rule of Taiwan.¹⁵³ In addition, the power asymmetry between China and Taiwan was far more serious than that between North and South Korea. It is fair to assess, therefore, that the security situation facing Taiwan was even worse than that of South Korea, and Taiwan desperately needed to improve it.

Taiwan’s frustration with the terms of the alliance was displayed soon after its conclusion. Toward the end of the First Taiwan Strait Crisis, the U.S. government requested that the Taiwanese forces withdraw from two offshore islands, Quemoy and Matsu, given the fear of logistical difficulties in the case of Chinese attack. Chiang adamantly refused to comply, however, and insisted that Taiwan would risk a direct clash with China even without U.S. support.¹⁵⁴ After the crisis de-escalated without a major incident, Chiang decided to deploy even more troops on these two islands, the decision that Dulles later called “foolish” when the Second Taiwan Strait Crisis occurred in the summer of 1958.¹⁵⁵ Even then, Taiwan continued to demand that the United States commit itself to defending the islands while refusing to reduce the Nationalist Chinese troops there.

Among the three U.S. allies, Taiwan was the most earnest proponent of a multilateral security system in East Asia. In October 1953, the Taiwanese government presented to the South Korean government a draft of a mutual defense pact, which it hoped would later expand to include other states in the Pacific region.¹⁵⁶ In 1954, Taiwan closely coordinated with South Korea to establish APACL. Taiwan also expressed a great interest in joining SEATO and maintained its interest in joining the organization even after its membership had been denied initially.¹⁵⁷ Taiwan also showed an interest in creating a collective defense system among South Korea, South Vietnam, and itself. For instance, a policy memorandum, dated March 16, 1956, states that “the positions of China, Korea, and Vietnam, who are vanguards against Communism, are more or less identical. It is necessary and possible that these three countries be united in an alliance.”¹⁵⁸

Taiwan was also willing to commit its troops to contribute to regional security. When the Korean War broke out in 1950, Chiang Kai-shek offered to send 33,000 troops to the Korean Peninsula, although the offer was declined by the United States, which feared that accepting it would lead to an expansion of the fighting.¹⁵⁹ This did not stop Taiwan from seeking to support South Korea during the war; the Taiwanese government helped South Korea and the UN Command in Korea to conduct psychological warfare and encourage the communist POWs to “repatriate” to Taiwan.¹⁶⁰ Taiwan also indicated several times its willingness to aid the South Vietnamese militarily during the Vietnam War.¹⁶¹

It is also noteworthy that Taiwan stepped up its efforts to create a multilateral security system when it perceived the weakening of U.S. security commitments. Around the time the Kennedy administration came to power in 1961, its high-ranking officials made a series of comments that signaled that the new U.S. administration was contemplating improving relations with China.¹⁶² Alarmed by the prospect of weakening U.S. security commitments, the Taiwanese government launched initiatives to weave the United States into a regional multilateral security system. In January 1961, Taiwan collaborated with the Philippines, South Korea, and South Vietnam to convene the foreign ministers conference, where it sought to lay the foundation for a regional security mechanism under U.S. leadership.¹⁶³ In May 1961, during Vice President Lyndon Johnson’s visit to Taipei, Chiang proposed the formation of a NATOlike organization among the United States and its allies.¹⁶⁴ From late July to early August, Vice Premier Chen Chen met with President Kennedy and proposed the creation of a joint multilateral military organization under the U.S. commander in chief in the Pacific. This organization would facilitate intelligence sharing and military training among U.S. allies under the leadership of the United States.¹⁶⁵ On August 15, Taiwan’s ambassador to the United States, George Yeh, handed to U.S. Secretary of State Dean Rusk “An Outline Proposal for a Collective Security Organization of the Anti-Communist Countries in the Western Pacific Asia.” According to the proposal, Taiwan, South Korea, Thailand, South Vietnam, and the Philippines would contribute troops to create a joint land force under U.S. command, while the United States would provide naval and air power and logistical support.¹⁶⁶

Unlike South Korea, Taiwan demonstrated its desire to include Japan in a regional collective security system. Taiwanese archival documents indicate that its government considered the inclusion of Japan to be crucial for creating a viable regional alliance.¹⁶⁷ When Chiang Kai-shek met U.S. Vice President Richard Nixon in November 1953, he argued that “a pact between the Republic of China, Japan, and the Republic of Korea [was] necessary,” and elaborated that he took this position despite the damage that his country suffered during the war with Japan.¹⁶⁸ Chiang accordingly tried to mediate between South Korea and Japan and to persuade South Korean President Rhee to admit Japan into a multilateral alliance in Asia.¹⁶⁹ Given the difficulty in persuading Rhee, Chiang even proposed in his meeting with Nixon that Taiwan could sign two bilateral security treaties, one with Japan and the other with South Korea, in order to bypass Rhee’s anti-Japanese stance.¹⁷⁰

Why did Taiwan take a more pragmatic attitude toward Japan than South Korea, given that both had suffered under Japanese imperialism and desired offensive actions against the communist bloc? An answer lies in the security environment with which Taiwan had to cope. As discussed above, even after the conclusion of the U.S.-Taiwan alliance, U.S. commitments to Taiwan were not firm, as symbolized by the lukewarm U.S. attitude toward the defense of the offshore islands. That the United States began the so-called Ambassadorial Talks with China after the end of the First Taiwan Strait Crisis amplified Taiwan’s concern that Washington might soon abandon Taiwan and reach out to China.¹⁷¹ Given these factors, Taiwan felt it more urgent than South Korea did to boost its security by joining a multilateral security system. This preference was reflected in the comment made by Ambassador Koo, who stated in July 1954 that “of course if Free China were included in a multilateral security pact, this would be even better than a bilateral pact.”¹⁷²

The geopolitical value of Taiwan for the United States had increased since the breakout of the Korean War. NSC 166/1, adopted in November 1953, stated that “the Chinese government on Taiwan [was] a considerable asset to the U.S. position in the Far East” and that “the military forces of the Nationalists constitute[d] the only readily available strategic reserve in the Far East.”¹⁷³ In 1953, Taiwan possessed almost 500,000 troops including an army of 422,000.¹⁷⁴ In addition, unlike South Korea, Taiwan is surrounded by water, and its homeland defense could be accomplished with a relatively small force. Hence, the U.S. government evaluated that Taiwanese forces “constitute[d] the only visible source of manpower,” should Chinese interventions in Indochina or Korea “make large-scale U.S. action against China necessary.”¹⁷⁵

U.S. policy documents reveal that one of the factors that enhanced Taiwan’s geopolitical significance for the United States was Japan’s reluctance to rearm.¹⁷⁶ In one paragraph before emphasizing Taiwan’s importance, NSC 166/1 expressed a pessimistic view of Japan’s military capabilities. It states, “There are no immediate prospects of rapid development of strength in the two countries which, potentially, can contribute most to a restoration of balance of power in Asia—Japan and India,” and “it will be some time even under optimum conditions, before Japan possesses the capability of exercising leadership in Asia.”¹⁷⁷ Given these conditions, the U.S. government decided to keep sending its military aid to Taiwan “under continuing review in light of the development of Japanese forces.”¹⁷⁸ The U.S. assistance for the development of Taiwanese forces continued in the second half of the 1950s; NSC 5723, adopted in October 1957, stipulates that one of the U.S. goals toward Taiwan was to help it develop its military “to contribute to collective non-Communist strength in the Far East.” The document also notes that the United States should “encourage conditions which [would] make possible the inclusion of the GRC in a Western Pacific collective defense arrangement.”¹⁷⁹

Table 1 shows whether and to what extent the evidence from the case studies fits the existing explanations for East Asia’s hub-and-spokes alliance system. First, the United States’ preference and pursuit of a multilateral security alliance strongly contradict the powerplay and collective identity hypotheses that emphasize the U.S. intention to create the hub-and-spokes system. As NSC 5429/5 and other documents reveal, the Eisenhower administration considered that a defensive multilateral alliance would serve U.S. interests best because it would reduce U.S. security burdens while maintaining stability in East Asia (table 2). The powerplay hypothesis also fails to explain Japan’s behavior; despite its prediction that Japan, as a weaker ally, should pursue a multilateral alliance which could more effectively constrain the United States than a bilateral one, Tokyo’s preference proved to be the complete opposite. The historical memory hypothesis, although it could explain South Korea’s refusal to include Japan in a multilateral alliance, cannot explain why Taiwan and the United States willingly established and tried to enhance security ties with Japan. The regional norm hypothesis explains none of the cases of the three U.S. allies, which were willing to accept significant limitations on their sovereignty in return for formal U.S. alliances.

In contrast with the existing explanations, the hypotheses derived from the social exchange network approach match the evidence from the case studies. As expected by hub’s preference hypothesis, the United States initially prioritized and sought an alliance with Japan, whose geostrategic values and potential capabilities were more highly evaluated than those of South Korea or Taiwan, while being reluctant to do so with South Korea and Taiwan initially. Washington did so not merely to form an alliance with Japan, but also to encourage Japan to contribute to regional security in East Asia. In the U.S. view, the ideal security architecture was one in which Japan, closely aligned with the United States, played a central role in contributing to regional security while the United States provided supplemental support for South Korea and Taiwan (figure 1A). Japan’s contributions to regional security, however, proved much less than Washington had wished. Tokyo’s insufficient security contributions prompted Washington to reevaluate South Korea’s and Taiwan’s geopolitical significance, reinforcing its bilateral ties with both. This policy change by the United States is consistent with the hub-spoke negative connection hypothesis.

The non-U.S. cases support the spoke’s preference hypothesis that the U.S. allies prioritized security ties with the United States rather than those among themselves. Japan sought and clung to a bilateral alliance with the United States more persistently than the United States did. South Korea and Taiwan clearly regarded security ties with the United States as vital for their survival. In fact, obtaining a formal alliance with the United States was so important that South Korea resorted to coercive binding—releasing the communist POWs and threatening to withdraw from the UN command—when Washington remained reluctant to formalize an alliance with Seoul at the end of the Korean War. Seoul’s behavior fits the coercive binding hypothesis, which provides a more complete explanation for the formation of the U.S.-South Korea alliance than the powerplay hypothesis that only emphasizes the U.S. efforts to prevent Seoul’s unilateral actions. The Taiwan case, though not explicitly supporting the coercive binding hypothesis, provides indirect confirming evidence for the hypothesis. Because China’s initiation of the First Taiwan Strait Crisis prompted Washington to formalize an alliance with Taiwan, Taipei needed not use coercive binding. However, Taiwan’s willingness to accept risk, such as its defiance of the repeated U.S. requests to withdraw or reduce its troops from the offshore islands, can be interpreted as a way to prevent the further weakening of U.S. alliance commitments.

The three non-U.S. cases also support the inter-spoke negative connection hypothesis, which posits that the degree to which the U.S. allies pursue security ties among themselves is negatively connected with the strength of their respective security ties with the United States. Japan succeeded in obtaining strong U.S. security commitments by and large, and its sense of security, further boosted by the U.S. signing of bilateral alliances with South Korea and Taiwan, minimized Japan’s incentives to develop security ties with South Korea and Taiwan. Taiwan remained most dissatisfied with the level of U.S. security commitments. To address its insecurity, Taiwan became the strongest advocate of a multilateral alliance among the three U.S. allies, reflected by its efforts to include Japan in such a system despite suffering from Japan’s past imperialism. Although its most desirable option was an offensive multilateral alliance, even a defensive one would have served its interests by stabilizing U.S. security commitments (table 2). South Korea’s pursuit of a multilateral alliance excluding Japan also fits this hypothesis. Given that the alliance with the United States ensured South Korea’s survival, Seoul sought an offensive rather than defensive multilateral alliance (table 2), to which Japan was an impediment. Seoul also feared that strengthening security ties with Japan might lead Washington to disengage from South Korea.

All in all, why did the United States, despite its overwhelming power, fail to create its preferred multilateral alliance (figure 1A), resulting in the persistence of the hub-and-spokes alliance system (figure 1B)? One impediment to a defensive multilateral alliance desired by the United States was Japan’s unwillingness to provide security regionally. Although the United States succeeded in creating an alliance with Japan, its strong security commitments to the ally, plus its alignments with the other two “spokes,” ironically weakened Tokyo’s incentives to develop security ties with Seoul and Taipei. South Korea’s refusal to include Japan in a multilateral system made the creation of the U.S. desired system even more difficult. Taiwan was the most earnest advocate of a multilateral alliance, but its efforts to create one were inconsequential given Japan’s and South Korea’s policies stated above. Under such circumstances, the United States had to make the most of the existing alliances and upgraded the role of South Korea and Taiwan, further stabilizing the hub-and-spokes system. Considering these observations, the hub-and-spokes system was an unintended consequence produced through the interactions among the United States and its three allies.

In this article, I explored why East Asia’s hub-and-spokes alliance system has persisted. In so doing, I pointed out the inadequacies of the existing explanations and proposed the social exchange network approach to explain the puzzle. The empirical analysis showed that this approach can effectively explain the preferences and behavior of the United States and its South Korean, Taiwanese, and Japanese allies, and how the interactions among them generated the hub-and-spokes system.

If this study is any guide to the origin of East Asia’s hub-and-spokes alliance system, what does it indicate for East Asia’s future alliance system? What are the impacts of China’s rise and the United States’ relative decline on the hub-and-spokes system? On the one hand, these trends have been influencing the existing system in a way that the United States had originally desired. Fearing the relative decline in U.S. capabilities to provide security, Japan has become more willing to contribute to regional security. After scuttling numerous self-imposed restrictions on its military policies, the Japanese government changed the long-held interpretation of Japan’s constitution in July 2014, making it possible to use force to contribute to U.S. military actions even when Japan is not under direct attack.¹⁸⁰ Japan has also been upgrading its security ties with other “spokes” by intensifying maritime security assistance for several Southeast Asian states, by upgrading security ties with Australia, and seeking to do the same with India.¹⁸¹ With South Korea, with which Japan continues to have seriously troubled relations resulting from historical issues, Japan remains willing to maintain security cooperation.¹⁸² Even on the question of Taiwan, Japanese leaders have become more willing to play a role in boosting the island nation’s security against China.¹⁸³ The leadership that Japan demonstrated in finalizing the Comprehensive and Progressive Trans-Pacific Partnership after the U.S. withdrawal from its original agreement was unthinkable during the Cold War.¹⁸⁴ Although Japan can certainly do more, Japan experts regard these changes in Japan’s security policy as a departure from its “pacifist” past.¹⁸⁵

On the other hand, China’s rise has dramatically increased not only its threats but also its capability to provide benefits for others, and such a capability has caused, and will continue to cause, complexities for U.S. relations with some other “spokes,” such as South Korea and Taiwan. Although South Korea still benefits from U.S. security commitments, it increasingly believes that it also needs better security ties with China to control North Korea. This security requirement makes South Korea’s position between the United States and China more ambivalent.¹⁸⁶ In the case of Taiwan, the problem may lie not so much with Taipei as with Washington. Public opinion polls show that despite significant economic dependence on China, Taiwan retains its strong preference for the status quo or even a more distinctly Taiwanese identity.¹⁸⁷ Because of the increasing estimated cost of defending Taiwan against China, however, U.S. willingness to provide security for Taiwan has been weakening. Some elites have begun to support disengaging from Taiwan.¹⁸⁸ A recent survey by the Chicago Council on Global Affairs also shows that 59 percent of Americans oppose U.S. military intervention if China attacks Taiwan.¹⁸⁹ U.S. security ties with these two allies may be further weakened if China effectively uses wedge strategies.¹⁹⁰

Under such circumstances, Washington must take two steps if it desires to preserve or facilitate the evolution of the hub-and-spokes system in a way that is conducive to its interests. First, it needs to retain its capabilities to provide security for its allies to ensure their desire to continue security exchanges with the United States. Although pessimists are quick to argue that China will be a hegemon in East Asia in the near future, the United States still retains, and should be able to maintain, the combination of resources that enables it to provide security for its allies if it adjusts its policy to face China’s increasing anti-access, area denial capabilities.¹⁹¹ Instead, if Washington stops doing so and tries to strike a “grand bargain” with China, the U.S. allies, South Korea and Taiwan in particular, are more likely to gravitate toward China.¹⁹²

Second, the United States should encourage its allies to strengthen security ties among themselves, preferably with “carrots” but with “sticks” if necessary. Such efforts must be aimed especially for facilitating security cooperation between South Korea and Japan. As this study shows, mistrust between Tokyo and Seoul made the U.S. goal of creating a multilateral alliance in East Asia extremely difficult to achieve, and the relations between them have become even more hostile, as historical issues have recently heightened mutual antagonism. This study also shows, however, that, in the past, U.S. security ties with its allies weakened their incentives to reconcile their past animosities. Now that both Seoul and Tokyo have begun to doubt the durability of U.S. security guarantees, the United States can more credibly use the threat of reducing its security commitments to persuade them to cooperate with each other.

The author thanks the anonymous reviewers for their insightful comments. He is also grateful to the following individuals for their helpful comments on the earlier versions of this article: Nicholas Anderson, Victor Cha, Timothy Crawford, Yong Deng, Iain Henry, Sean Lynn-Jones, Tongfi Kim, Evan Resnick, Yuan-kang Wang, David Wolff, and Toshio Yamagishi. And for support and advice regarding archival research in Japan, South Korea, Taiwan, and the United States, he thanks Masaya Inoue, Ayako Kusunoki, Jong-won Lee, Haruka Matsumoto, and Shingo Yoshida. The research for this article was partially funded by Grants-in-Aid for Scientific Research, provided by the Japan Society for the Promotion of Science (No. 19H00575 and No. 15K03333).