Introduction
In the summer of 2010, the United States and Russia were involved in a brief spying scandal. Despite the arrests, criminal convictions, and eventual exchange of the alleged spies, each side immediately played down concerns that this scandal might negatively influence their relationship. Russia's Prime Minister Vladimir Putin quickly blamed the whole affair on some overzealous US police officers. Similarly, the US Assistant Secretary of State, Philip Gordon, dismissed the Cold War flashbacks, claiming that the two states were in fact, ‘moving towards a more trusting relationship’.Footnote 1
This mixture of action and discourse that points at the same time to the presence and absence of trust raises serious questions for scholars of international politics. On the one hand, rounding up a group of individuals and accusing them of spying for another country is precisely the type of action that suggests a lack of trust in others. On the other hand, the official statements stress the progress towards a trusting relationship between these two states. Given the contradictory interpretations of such events, how could an observer tell which one is correct? In short, how would we know whether a trusting relationship between these two countries exists?
In the past fifteen years, International Relations has generated significant scholarship grappling with the theoretical aspects of trust and using it to explain empirical phenomena.Footnote 2 Despite these contributions, we argue that they fail to answer convincingly a key question central to our opening illustration: ‘How can one identify a trusting relationship between two states?’Footnote 3 This question is fundamental if the concept of trust is to play a role in both study and practice of international relations. It must logically come prior to questions such as ‘what role did trust play in establishing that relationship?, or ‘how can trust be built in relations among states?’Footnote 4 Though these latter questions are important in their own right, there first has to be some way to identify that what is being examined is indeed a trusting relationship.
We identify three prevailing ways of determining the existence of a trusting relationship in the current scholarship: the presence of cooperation, discursive expressions of trust, or the voluntary assumption of vulnerability. Each of these, however, has shortcomings. With regard to the first approach, we demonstrate that it is insufficient to take the cooperation between states as a sign of a trusting relationship. Similarly, we show that accepting someone's word as a proof of the existence of a trusting relationship has serious limits. Third, we argue that the fact that one or both sides have voluntarily taken on vulnerability does not mean that their relationship is trusting. Instead, we propose that focusing on how states decline to adopt or remove existing hedging strategies provides a better alternative to determining trusting relationships in international politics.
The ability to identify trusting relationships matters because such relationships show that states have the ability to transcend the security dilemma.Footnote 5 Trusting relationships suggest the potential for transformation in political relations even under the conditions of international anarchy. Our approach, based on the degree to which actors adopt hedging strategies, creates a high empirical threshold for the existence of a trusting relationship compared to other methods, but the payoffs resulting from it are potentially substantial. It allows scholars to focus on long-term historical processes between states at the international level. It can be used to demonstrate how trusting relationships can form and continue to exist between states independently of whether individual political leaders are favourably or negatively predisposed to trust. Finally, our approach is also able to work in reverse to analyse the onset and the development of a distrusting relationship.
To substantiate our claims, the article proceeds in three parts. First, we outline a crucial conceptual distinction between trust and confidence to establish the theoretical starting point for our argument.Footnote 6 Second, we will argue for the usefulness of employing the level of hedging among actors as an analytical tool to detect trusting relationships. Third, we show how our approach can be applied to empirical illustrations. Using the strategic relationship between the United States and the Soviet Union during the Cold War, we argue that the existing methods provide ambiguous results, allowing for the counterintuitive possibility that this was a trusting relationship. We then demonstrate the utility of our approach by tracing the development of a trusting relationship in the security alliance between the United States and Japan during the Cold War.
Trust, confidence, and trust-as-confidence
We conceive of trust as the ideational structure that cognitively reduces or eliminates the residual risk and uncertainty that is part of any decision. This definition draws on work by Niklas Luhmann,Footnote 7 who argued that the function of trust is to allow actors to cognitively reduce or eliminate the overall amount of risk and uncertainty they face in making decisions. The presence of trust leads actors to conduct themselves as if there were a greatly reduced or non-existent possibility of unwanted outcomes in their relationship with another actor.
Building on this, we define a trusting relationship as an intersubjective ideational structure that allows two or more actors to partially or wholly set aside existing risk and uncertainty in the actions of others in favour of an assumption that the accepted social norm or particular agreement between the two actors will hold. Our emphasis on trust being particular to a specific social norm or agreement follows some scholars who argued that there can be no generalised trust between actors.Footnote 8 While we do not dismiss the idea that generalised trust can exist, even if it does, trust will likely vary on particular issues. It is analytically more useful, we believe, to discuss trust on the basis that X trusts Y to do Z, instead of claiming that X trusts Y in general. This is particularly the case in the complex relationships between states, in which the degree of trust might vary across issues.
There are three primary benefits of trusting relationships in international politics. First, they simplify the complex reality of international anarchy by reducing it to ‘manageable proportions’ by allowing ‘most of the contingently possible future events … [to be] thought of as zero for all practical purposes’.Footnote 9 Second, they allow actors to sustain agreements that are much more stable than would otherwise be expected under conditions of anarchy.Footnote 10 Third, trusting relationships enable actors to avoid the costs and negative structural consequences linked to the alternative response, which is to hedge against the risk.Footnote 11
In contrast to trust, we define confidence as a calculation based on an actor's perceptions that another actor will reciprocate a costly signal or act according to a predetermined agreement or social norm. So conceived, confidence is a straightforward rational choice calculation. The fundamental difference between confidence and trust is that trust is an intersubjective social structure that cognitively reduces or eliminates the overall amount of risk and uncertainty that actor faces, while confidence is simply the calculation that an actor will reciprocate cooperation or follow an agreement or social norm. Confidence does not reduce the perception of risk, trust does. The problem with some trust scholarship is that it limits the concept of trust to such calculations, conflating confidence with trust.Footnote 12 We call this approach trust-as-confidence in order to clearly differentiate it from our conceptualisations of confidence and trust.Footnote 13
Philosophically, the most famous version of trust-as-confidence was proposed by Russell Hardin, who argued that trust is encapsulated interest, or a situation where ‘I trust you because I think it is in your interest to take my interests in the relevant matter seriously.’Footnote 14 This envisages a specific type of a trusting relationship based on a rational choice calculation. Such a conceptualisation is echoed in the field of International Relations most prominently by Andrew Kydd. He holds trust to be a rational decision-making process determined by the probability that the other actor will reciprocate cooperation.Footnote 15 Even in what purports to be alternatives to Kydd's approach, formulated in the works of Aaron Hoffman, Ken Booth, and Nicholas Wheeler, or Brian Rathbun, there are similar underlying assumptions. While they stress the normative underpinnings of trusting relationships, these authors nevertheless fall back on a conceptualisation of trust based on actors making predictions about other actors' future behaviour. What distinguishes them from Kydd is the inclusion of the moral value of the trusting relationship in their estimate of these future expectations. For Hoffman it is the belief that the other state will not abuse the power which they hold over the truster, because the two parties have ‘chang[ed] the rules, norms, and principles (‘institutions’) that govern their relations’.Footnote 16 Booth and Wheeler, alternatively, highlight the importance of interpersonal dynamics between particular leaders in overcoming security dilemmas.Footnote 17 Finally, Rathbun accounts for trusting relationships by the presence of generalised trust, which ‘rests on a general belief in the benevolent character of others’.Footnote 18 Despite their contribution to the understanding of trust in international politics, all of these approaches still conflate trust and confidence. This is because they conceive of trust solely as an additional factor in actors' calculations about the future. In short, they have no conceptual space to appreciate a trusting relationship as an intersubjective ideational structure, which allows actors to cognitively reduce the risk and uncertainty they face.
Undoubtedly, in conventional practice the words ‘trust’ and ‘confidence’ are often used interchangeably, which might explain why so many scholars have taken the route of studying trust-as-confidence.Footnote 19 The problem with trust-as-confidence is that it completely removes the ideational nature of trust that might make it a distinct social phenomenon. In other words, it does away with the social nature of trust. As Lewis and Weigert put it, ‘these researchers are not really studying trust at all. What they are investigating are the processes by which individuals come to formulate and act on the predictions about the behaviour of others’.Footnote 20 Therefore, for the concept of trust to be useful analytically within International Relations, it needs to be distinguished from trust-as-confidence. This is necessary to preserve the most significant attribute of trust, the cognitive reduction of risk.
Identifying trusting relationships
The contribution of distinguishing between trust and confidence becomes obvious when compared to the existing ways of identifying trusting relationships in international politics. There are currently three main approaches: cooperation, trusting discourses, and the voluntary acceptance of vulnerability.
The first method identifies trusting relationships by looking for cooperation. This method can be developed through the extension of Andrew Kydd's work. Although Kydd does not develop the concept of trusting relationship in any of his writings, we argue that his scholarship points in the direction where the incidence of cooperation signifies the presence of a trusting relationship. In the absence of cooperation, there would be no way to tell whether actors held the belief that the other side was trustworthy. This is because Kydd defined trust as a ‘belief that the other side is trustworthy, that is, willing to reciprocate cooperation, and mistrust as a belief that the other side is untrustworthy, or prefers to exploit one's cooperation’.Footnote 21 He further claimed that, ‘cooperation requires a certain degree of trust between states’,Footnote 22 and that ‘when we see conflict it is a sign that one or both of the states are likely to be untrustworthy’.Footnote 23 This suggests that where sustained cooperation exists, trust must also be present; else the observed cooperation would be impossible.Footnote 24 As Kydd claimed, ‘mutual trust is necessary for cooperation’.Footnote 25
Whereas this is perhaps an intuitive deduction, it suffers from two basic problems. First, cooperative behaviour can be a result of rational payoffs or, in our nomenclature, confidence. Trust, understood in its social dimension, is not needed to explain the behaviour. Should the expected value of a decision to cooperate be greater than existing alternatives, states might choose to cooperate even with the real risk of defection or cheating, since this potential behaviour would be simply factored into their calculations. Second, a lack of cooperation does not necessarily indicate the absence of a cooperative mentality or distrust. There could simply be little to gain from cooperation given other opportunities. As David Good summed up, ‘while cooperation and trust are intimately related in that the former is a central manifestation of the latter, the former cannot provide, for either the actor or the analyst, a simple redefinition of trust’.Footnote 26
Another technique to determine a trusting relationship in international politics is to examine public and private statements or interview political decision-makers, using their discourse to decide whether they trusted a particular person or institution.Footnote 27 Though again this might be an intuitive way to tackle the issue, this approach is problematic for two reasons. First, there is a problem with interpreting what would almost certainly be strategic language.Footnote 28 Instead of providing clear access to their beliefs, actors are more likely to give answers or to write accounts, even in private papers, that suit their political purposes. Furthermore, as Luhmann contends,
Although the one who trusts is never at a loss for reasons and is quite capable of giving an account of why he shows trust in this or that case, the point of such reasons is really to uphold his self-respect and justify him socially. They prevent him from appearing to himself and others as a fool, as an inexperienced man ill-adapted to life, in the event of his trust being abused.Footnote 29
Not only will such statements be shaped by political expediency, but also by the desire of actors to uphold their image. One can expect that actors will tend to portray successful trusting actions as acts of trust, whereas unsuccessful trusting endeavours will not be characterised as such – to avoid looking foolish.
Finally, scholars argue that a trusting relationship can be identified when an actor voluntarily takes on vulnerability.Footnote 30 The problem with this approach is that an actor in a trusting relationship is unaware of their vulnerability.Footnote 31 The voluntary acceptance of vulnerability and being in a trusting relationship are actually opposite conditions. Actors in a trusting relationship will have little to no feeling of vulnerability precisely because trust functions to cognitively reduce or eliminate their perception of risk in the situation. To state that a trusting relationship can be identified by observing actors voluntarily accepting vulnerability cannot be correct if trust has the function of cognitively reducing risk and uncertainty.
This focus on the voluntary assumption of vulnerability could arise from scholars conflating trusting relationships and strategies to enter into a trusting relationship. While we make no claims in this article as to whether an actor's ‘willingness to be vulnerable’ might be a useful strategy to gain trustFootnote 32 and thus enter into a trusting relationship, we believe that such felt vulnerability from the perspective of the actor cannot signal the existence of a trusting relationship – indeed, it signals the opposite.
The increased vulnerability might be something observed by the analyst as an indicator of a trusting relationship, but the existence of a trusting relationship means that vulnerability is not knowingly experienced by the actor. In the words of Nicholas Rengger, ‘“trusting” is not done “consciously,” as a matter of considered reflection, … [it is] a matter of habit’.Footnote 33 Actors might find themselves in a relationship that involves vulnerability, which is readily apparent to the observer, but trust allows them to be free ‘from worry and from the need to monitor the other's behavior or to extricate herself from the relationship’.Footnote 34 Our point here is that the voluntary acceptance of vulnerability and being in a trusting relationship are incompatible from the actor's perspective.
Trust, confidence, and hedging strategies
In his article ‘A Conceptualization of Trust in International Relations’, Aaron Hoffman suggested a number of strategies that might be employed to detect a trusting relationship in international politics. The first strategy examined the perceptions of the leaders involved in a decision and connected these perceptions to their subsequent policy choices. When leaders, in their relations with other states, voluntarily put themselves in a position of vulnerability because they believe the other states to be trustworthy, Hoffman considered this to be evidence of a trusting relationship. The second strategy examined the mechanisms selected by each state to monitor cheating. Here, Hoffman argued that where oversight occurs before actions are taken there is a less trusting relationship than in cases where oversight occurs after the consequences of an action are revealed. The third strategy examined the restrictiveness of rules that structure the cooperation between two states. The more leeway granted by the agreed-upon rules, the greater the likelihood that a trusting relationship exists.Footnote 35
Hoffman's article was clearly a major step forward for studying trusting relationships in international politics empirically. However, it maintains that a voluntary acceptance of vulnerability is necessary, which we claim is at odds with a social theory of trust. Additionally, though we do not take issue with his other two strategies per se, our approach generalises their logic to create a more comprehensive theoretical framework.
Our differentiation between confidence and trust allows scholars to analyse distinctive types of behaviour and determine the kind of relationship that states find themselves in. We assume that all relationships in international politics will continually be under pressure from the perpetual risk and uncertainty generated by the state of anarchy. Because of this potential for catastrophic outcomes, some scholars have even argued that there is ‘little room for trust among states because a state may be unable to recover if its trust is betrayed’.Footnote 36 Without an adequate way to assess the existence of a trusting relationship, there is no way to tell whether this scepticism is warranted or not.
In the absence of a trusting relationship that cognitively reduces uncertainty, it follows that states have to engage in hedging strategies.Footnote 37 Adopting a hedge reduces one's dependence on and vulnerability to others, for instance ‘by cultivating alternative partners, projects, and networks’.Footnote 38 Under the conditions of anarchy, hedging strategies allow states to self-insure against possible defection or opportunism by other states, allowing them to act more securely in a risky environment because the possible ‘worse-case’ outcomes are both anticipated and accounted for.Footnote 39 A standard example of a hedge in a military alliance could be the decision to keep more forces than are necessary should a partner renege on the agreement when its assistance is needed.Footnote 40
Hedging permits actors to lower their overall risk position, but it also creates problems. It may perpetuate self-reliance where more risky but rewarding cooperative behaviour is possible. Hedging may also contribute to a false sense of security and increased risk-taking behaviour, if the actors become overconfident about their hedge. Finally, hedging leads actors to focus on what might go wrong, pre-empting more optimistic strategies by perpetuating worst case scenario thinking.Footnote 41 Although the hedge protects states against the potential downside of defection, it also limits the gains should the partner be trustworthy. In sum, the hedge avoids the worst outcome and therein reduces the risk, but it comes with a cost that lowers the expected value of the position and potentially promotes structural logics with negative consequences.
Alternatively, if trust is the ideational structure that cognitively reduces or eliminates risk, then the presence of a trusting relationship will be indicated by the reduction, or even elimination, of hedging strategies. Actors will simply behave in a manner suggesting that there is no risk of defection, or at least significantly less than might otherwise be expected. Trusting relationships should therefore be marked by a large reduction or even elimination of hedging. Moreover, the link between hedging and trusting relationships holds independent of the level of risk faced by the actor.Footnote 42
Trust and hedging both cognitively reduce risk in the international system. Hedging cognitively lowers risk because it is assumed that taking action actually reduces risk in reality. Trust cognitively lowers risk independent of a reduction of the ‘real’ risk. Whereas hedging lowers risk but usually leads to the reduction in the expected value of the position and problematic behavioural outcomes, a trusting relationship lowers risk without a similar trade-off. Fundamentally, this is the primary reason why trusting relationships, where possible, are to be valued.
To determine whether a relationship between two states is trusting, an analyst can use two methods. One approach, building upon Hoffman, is to create categories with different levels of hedging activity. The extent of hedging indicates whether a relationship is trusting. Hoffman's examples have an option of distrust that exhibits a cost, the hedge, and an option of trust that removes some hedging strategies. For instance, the less risky before-the-fact oversight requires more of an institutional framework, that is, hedging, than the more risky after-the-fact oversight. Essentially this method relies on a single point in time to ascertain the quality of the relationship.
The second approach offers a more generalised way to determine the presence of a trusting relationship. It looks at changes in hedging behaviour across time and decides that there is sufficient evidence to claim that a trusting relationship has been developing because otherwise the reduction of hedging strategies would be difficult to explain. Crucially, it is only through examining long-term behaviour that we can differentiate between the development of a trusting relationship and a potential short-term lack of hedging that might be used as a signalling device.
The focus on hedging is not designed to answer the question of how a trusting relationship came about. It is a method to recognise one. We argue, and empirically demonstrate in the following sections, that an unobservable social structure, a trusting relationship, has causal effects on policymaking. These effects can be observed through a material proxy, which is the presence or absence of a hedging strategy.Footnote 43 This proposition is helpful in answering our central question of determining a trusting relationship. We are, however, aware that the presence or absence of hedging is multicausal and cannot serve on its own as the evidence of a trusting relationship.
In order to identify a trusting relationship, an analyst needs to make two judgements on the situation. First, hedging, though material, cannot be understood outside of the social meaning of the hedge itself. As Wendt put it, ‘material resources only acquire meaning for human action through the structure of shared knowledge in which they are embedded’.Footnote 44 For example, US nuclear weapons, a clear hedge against potential attack, did not have the same meaning for the British during the Cold War as they did for the Soviet Union.Footnote 45 The analysis must take into consideration the historical understandings of the hedge – sacrificing an animal to the gods to increase the likelihood of victory in battle might be considered a hedge even if the analyst believes the hedge to be useless from their perspective. In general, the analyst must understand the purpose of the hedge from the perspective of the actors in order to identify that the hedge is indeed relevant to the relationship between the two states. Trusting relationships cannot be studied without some reflection of the perceptions of leaders;Footnote 46 however, we propose that the focus on the hedge provides a more reliable indicator of a trusting relationship than assertions of its existence by actors.
Second, taking into account the geographical and historical situatedness of the states in question, an analyst must make an assessment whether the state had an adequate hedge. This judgement must consider three factors. First, the presence or absence of hedging strategies will be determined by the amount of risk that a state is willing to live with, that is, its risk preferences, be they relatively averse, neutral, or taking. Our approach presupposes that hedging will only occur if states are risk averse; a presupposition that we argue is likely, though not necessarily universal.Footnote 47 Second, the absence or presence of hedging must be evaluated in the light of the state having the ability to hedge. There cannot be insurmountable material constraints to the prospective hedging action. Third, there needs to be some consideration whether the presence or absence of a hedge was caused by other factors, and whether the significance of these factors is great enough to influence the assessment. For instance, one might argue that the reason why states forgo some weapon systems is because there is a strong domestic opposition to military spending.
To conclude, our approach offers several advantages. First, it escapes the rational choice framework to provide a way of identifying trusting relationships based on a social theory of trust. Thus it does not run the risk of conflating cooperation and trust. Second, because there is an emphasis on the material nature of the hedge, it does not depend on political discourse that can reflect strategic language rather than ‘true’ beliefs and perceptions. Third, it operationalises the habitual nature of trust. It avoids the theoretical and empirical pitfalls of using voluntarily accepted vulnerability as an indicator of a trusting relationship. Finally, our approach provides a framework where a trusting relationship can be confirmed or refuted given the limits to epistemology in the social sciences. Importantly, this means our approach makes it possible to empirically support the realist argument that trusting relationships are rare or even do not exist, potentially demonstrating the irrelevance as much as relevance of trust in International Relations.
Having outlined our approach, we now will proceed with two illustrations. In the first illustration, we use our approach to show why the relationship between the United States and the Soviet Union with respect to nuclear aggression was not trusting. This might seem an obvious conclusion. The issue is that all three methods currently prevalent in International Relations lead the analyst to a false positive attribution of a trusting relationship. In the second illustration, we trace the development of trusting relationship in the military alliance between Japan and the United States to show how our method can be used to identify such developments.
False positives in the US-USSR nuclear relationship
The nuclear relationship between the United States and the Soviet Union highlights the weaknesses of the current approaches to the study of trust in International Relations. Despite evidence of cooperation between the superpowers manifested in nuclear treaties, the voluntary acceptance of vulnerability entailed in mutual negotiations and nuclear deterrence, and verbal expressions of trust, including the famous remark ‘trust, but verify’,Footnote 48 we argue that the nuclear relationship between the Soviet Union and the United States cannot be characterised as trusting. In other words, all three current methods potentially yield false positives for the presence of a trusting relationship between the United States and the Soviet Union. In contrast, our method unequivocally points to a relationship of confidence, because at no point was the hedge against attack, each state's nuclear deterrent, ever eliminated or reduced to a point below which the other state would not face unsustainable destruction should the other decide to initiate hostilities.Footnote 49
This relationship of confidence, as opposed to a trusting relationship, was quite literally reflected in nuclear strategy. The United States and the Soviet Union strived to ensure that the other had no doubt concerning the outcome of potential aggression. They each took steps to decrease the chance of misperception, to make calculations about each other's behaviour more reliable, and therein avoid inadvertent nuclear war.
Nuclear deterrence is based on ensuring that the other side is highly confident in two factors. The first is the high probability of a retaliatory nuclear strike.Footnote 50 In the US-Soviet relationship, this was reflected in the continual development of assured second-strike capabilities on both sides, including the introduction and improvement of hardened silos and submarine-launched ballistic missile systems.Footnote 51 The second requirement is to ensure that that the other side believes that such a retaliatory blow is likely. As Hans Morgenthau argued, ‘Deterrence has thus far worked only because there has remained in the minds of both sides a doubt as to whether the other side was really bluffing. Or, to put it the other way around, both sides were able to give the threat of nuclear war at least a certain measure of plausibility.’Footnote 52 Both the Soviets and the US had a declaratory strategy whose aim was to assure the other of the likelihood of a retaliatory attack. The central role of nuclear deterrence, most strongly exemplified in the doctrine of mutual assured destruction, was clear on the part of both the United States and the Soviet Union.Footnote 53 Both sides developed strategic plans to ‘win’ a nuclear war, further demonstrating their resolve to respond with nuclear weapons should the situation arise.Footnote 54
Having communicated that survivable nuclear forces were likely to retaliate to offensive measures taken by the other side, each side understood that any offensive move would lead to severe consequences. This is borne out in casualty estimates on both sides, which expanded rapidly with the increase in the arsenals.Footnote 55 Even prospects for a ‘limited’ nuclear exchange in the 1980s were expected to lead to the deaths of tens of millions of US and Soviet citizens.Footnote 56
These two factors suggest that the United States and the Soviet Union were in a relationship of confidence with respect to each other's potential use of nuclear weapons, despite instances of successful cooperation and discourses that suggested trust. Additionally, the nuclear relationship between the United States and the Soviet Union illustrates the problems associated with using the acceptance of vulnerability as an indicator of trust. Nuclear deterrence based on mutually assured destruction, the dominant nuclear strategy during the Cold War, meant that both sides had to voluntarily put themselves in a position of vulnerability with respect to each other's nuclear forces, as was best illustrated by the Anti-Ballistic Missile treaty signed in 1972. This focus on deterrence was not constant throughout the Cold War, but when alternative views surfaced their popularity was generally short-lived.
An objection could be made that deterrence itself was a structural condition that enforced vulnerability. If this is the case, then there can be no claim to ‘voluntary’ acceptance leading to a false positive. However, there is ample evidence that neither state was fully confident in the deterrence paradigm, suggesting that it did not exert a structural pressure that effectively eliminated agency. At the same time that both states signalled their intention and capability to strike should the other side attack, the potential devastation made possible by an unintentional nuclear war led to several arms control agreements that incorporated confidence-building measures.Footnote 57 These were intended to prevent poor decision-making that might otherwise arise through, ‘misinformation, insufficient information, the lack of the ability of a decision-maker to comprehend the information, disbelief, differing perspectives, and the propensity to see what one wants to see in any situation’.Footnote 58 Thus, both sides wanted to ensure that there was effective signalling of their retaliatory capabilities and intent. The severity of a nuclear destruction led to additional confidence-building measures to ensure that such a conflict would not occur accidentally. The extent of the nuclear arms race itself also suggests that neither side was fully confident in deterrence. The arms race continued well over and above the need for minimum deterrence, despite the fact that, as Kissinger argued, ‘there exist maximum limits, beyond which the accumulation of weapons loses its political value’.Footnote 59
The focus on agreements to ensure that miscommunication did not lead to inadvertent nuclear war and the stockpiling of nuclear weapons well beyond minimum deterrent value were both hedges against the possibility of nuclear deterrence failure. This demonstrates that the deterrence regime itself cannot be considered a trusting relationship. Both sides made themselves vulnerable to the nuclear attack of the other, but their willingness to be vulnerable does not suggest the existence of a trusting relationship. To sum up, despite the presence of cooperation, trusting discourses, and the voluntary acceptance of vulnerability, we argue that the presence of each side's nuclear hedge is the best indicator that trust was not placed in the other side's nuclear intentions.
The US-Japanese alliance as a trusting relationship
Whereas our first illustration provides an example of a relationship of confidence and demonstrates why the current methods to detect trusting relationships all ‘admitted an example they should not’,Footnote 60 the second illustration offers an example of how our method can reveal the development of a trusting relationship between two states. We examine the evolution of the security relationship between the United States and Japan in the second half of the twentieth century. For the study of trust at the international level, this is a hard case because the relationship grew from the open enmity of the World War II. Moreover, especially in the beginning of the relationship, it entailed a great deal of inequality. Finally, despite the improving quality of the relationship in the area of security, there was no lack of tension between the two countries in the economic sphere as Japanese exports made their way onto the US market and US companies derided Japanese protectionism. As two scholars put it, given these tensions, ‘it is striking that the security relations between the United States and Japan have grown increasingly intimate’.Footnote 61 Indeed, the case is a prime example of how a trusting relationship develops in a specific issue area (X trusts Y with regard to Z). Hedging strategies, though available to both sides, have over time been removed, which indicates the evolution of a trusting relationship better than the alternative approaches.
It would be easy to base our characterisation of the Japanese-US security relationship as trusting only on the grounds of repeated verbal assurances such as those contained in security treaties or mutual declarations.Footnote 62 For instance, in January 1972, President Richard Nixon and Prime Minister Eisaku Sato issued a joint declaration ‘emphasizing the importance of US-Japanese relations being founded on mutual trust and independence’.Footnote 63 While paying lip service to mutual trust, the statement came at a time when the US-Japanese relations were undergoing fundamental shifts. Japan in particular was recovering from the ‘Nixon shocks’, a series of actions by President Nixon (his trip to China in July 1971, an economic policy defending the dollar, and the imposition of tariffs on Japanese imports), which Japan perceived were aimed against itself and which undermined its position in North-East Asia.Footnote 64 One historian characterised the relationship during the Nixon presidency in the following way: ‘Having played on the “China threat” and the importance of close ties to the United States to assure both security and prosperity, the LDP felt whipsawed and betrayed by Nixon.’Footnote 65 Ten years later, following a meeting between President Ronald Reagan and Prime Minister Zenko Suzuki, the language of their joint communique hardly changed. Once again, both sides ‘reaffirmed their solidarity, friendship, and mutual trust’.Footnote 66 Just like during the Nixon era, the US-Japanese relationship continued to face a series of challenges, mainly as a result of persistent tensions in the economic sphere, the questioning of commitment to their mutual relationship on both sides, and the Soviet military build-up.Footnote 67 Both of these instances support our earlier reservations about relying on verbal utterances when it comes to determining the existence of trusting relationships. The significant gap between what was said and what was done calls into question methods relying on such statements, be they public or private.
Similarly, because of frequent contestation in the economic sphere, the relationship between the two countries cannot be characterised as trusting simply on the basis of instances of cooperation. Throughout the Cold War, cooperation between Japan and the United States was repeatedly questioned and challenged. Both have at times accused each other of adopting hostile economic policies. There are numerous accounts highlighting the strains and tensions to which such cooperative arrangements were repeatedly exposed.Footnote 68 Given that the relationship generated cooperation as well as persistent disputes about it, it is impossible to decide whether cooperation in this instance indicates a trusting relationship or not.
More reliable evidence of the development of a trusting relationship rests in the abandonment and/or refusal of adopting hedging strategies by both countries vis-à-vis each other. Following the post-World War II occupation, the United States incrementally relinquished an unprecedented degree of control over Japan. It moved away from the view, frequently articulated in the early postwar years, that Japan could not be trusted as an ally unless there was a sizeable American contingent present in the country.Footnote 69 The history of the US military presence in Japan is a history of continuous withdrawal. The number of US soldiers declined from a high of over 210,000 in 1954 to roughly 84,000 a decade later, with further significant cuts in the early 1970s that stabilised the numbers at slightly below 50,000 for the rest of the Cold War.Footnote 70 While the United States clearly maintained a sufficient safety net in the form of its own overall military capability to counter any potential U-turns in Japanese foreign and security policy, it was gradually doing away with this hedging strategy. Additionally, the US withdrawals were typically accompanied by calls for greater Japanese military spending.Footnote 71 In the absence of a trusting relationship between the two states, such moves would hardly have made sense from the US point of view.
In mirror reflection to the US relinquishing its hedging strategies, there were a number of Japanese policies during the Cold War that disregarded or dismantled hedging strategies in the security relationship. Initially, Japan's military options were obviously heavily curtailed. The onset of the Korean War changed this and marked the beginning of the US demands that Japan spend more on defence. This prodding would not stop in the subsequent decades and, by the 1980s, turned into open frustration and even fairly aggressive demands. Despite the growing ability to spend more on their military and therefore adopting additional hedging strategies, Japanese defence spending stayed below 1 per cent of the GDP throughout much of the Cold War.Footnote 72 Even though Japan faced Soviet, Chinese, and to a lesser degree North Korean threats, it resisted increasing its spending on conventional forces. While in absolute terms this meant a steady climb in defence expenditure, up to a point where Japan became the third largest military spender in the world in 1989, the ratio to GDP was 3 to 4 times below that of other comparable states.Footnote 73
In addition to giving up or foregoing certain weapon systems, Japan also rejected the possibility of developing its own nuclear deterrent, preferring instead to remain under the US-extended nuclear deterrence. In effect, it was showing its trust in the alliance arrangement despite occasional US suggestions that Japan might consider obtaining its own nuclear hedge.Footnote 74 In the nuclear sphere the unwillingness to engage in explicit hedging strategies has given rise to a small industry of those either predicting or guessing the point when Japan, with its obvious technological capacities, would finally ‘go nuclear’.Footnote 75 However, by joining the Treaty on the Non-Proliferation of Nuclear Weapons, albeit after a lengthy domestic discussion, Japan formally foreclosed its own nuclear hedge.Footnote 76
While we are aware that there was significant domestic opposition to military spending in general and the acquisition of nuclear weapons in particular, these decisions cannot be entirely ascribed to domestic reasons. After all, Japan did benefit from the security alliance with the US and its nuclear umbrella, which raises an interesting counterfactual question about the strength of anti-militarism were these structures not in place and Japan had to fend for itself. Moreover, international pressures will occasionally lead decision-makers to ignore domestic constituencies. Notably, in our case, to preserve and strengthen the security relationship with the United States, successive Japanese governments did not hesitate to disregard the anti-nuclear domestic sentiments as well as their own public commitments not to manufacture, possess, or allow the entry of nuclear weapons to Japanese territory.Footnote 77 These two factors suggest that the lack of hedging cannot be sufficiently explained at the domestic level, and specific types of relationships with other states must be taken into account.
Although the Japanese behaviour could be characterised as free-riding or buck-passing,Footnote 78 and indeed had been derided as such especially in the US Congress, we argue that the unwillingness to adopt hedging strategies is better understood as a manifestation of a developing trusting relationship. Free-riding on US military power might have had an appeal in the 1950s and 1960s, because Japan was rebuilding after the World War II defeat, but it made little sense as the country became an economic powerhouse in the 1970s and 1980s. In an anarchic international realm long-term free-riding is a form of suicide. If alliances are as fragile as believed by realists, then free-riders put themselves at a major risk when their partners do not live up to their commitments. We should expect states to hedge against this possibility. When they do not, explanations for their unwillingness to do so must be sought elsewhere. This is exactly why trusting relationships, which cognitively reduce risk, must be part of the explanation for such behaviour.
To conclude, the US-Japanese relationship provides a good illustration of our argument about the development of a trusting relationship as a long-term historical process in which hedging strategies are continuously discarded and/or not adopted. What this indicates is that the existence of a trusting relationship is not a binary proposition. It is impossible to say that at one particular point the US-Japanese relationship became trusting. What our approach allows us to do, however, is to focus on hedging strategies or the absence thereof as behavioural manifestations of trust over a longer period of time.
Conclusion
This article has put forward an approach to determine the existence of trusting relationships based on the presence or absence of hedging strategies. We have argued theoretically and demonstrated empirically how this method can identify such relationships better than the three existing alternatives.
The approach may be criticised for raising the bar for trusting relationships in international politics too high because it is based on identifying specific actions, such as doing away with or foregoing hedging strategies, that states rarely engage in. However, we believe this high threshold to be justified for three reasons. First, the approach takes seriously the view that trusting relationships are relatively rare and difficult to achieve in international politics. Nevertheless, it does not mean that such relationships cannot exist. This is important, because the possibility of recognising the existence of trusting relationships suggests transformative potential in international politics. Second, our approach allows for, indeed, demands, an observation of long-term ideational as well as material processes, thus avoiding the pitfalls of taking actors at their own word or reading too much into short-term tactical policy shifts. Finally, the approach provides room for actually studying trust at the international level in relations among states. Trusting relationships are in this sense not a property of individual person-to-person dynamics, as evidenced by fortunate constellations like Reagan-Gorbachev, but rather a possibility, independent of particular leaders, in relations between collective units such as states.
