1.1 Interpreting Legitimization as a Social Cue

This Element develops a new theory to understand how international institutions shape the perceived legitimacy of foreign policy. It advances a social theory of legitimacy that builds upon political psychology, social identity theory, as well as insights about state-level norms from constructivist literature. It begins with the premise that legitimacy is in the eye of the beholder, so to theorize and uncover evidence about it, one must first begin with the audience of legitimacy: individuals.

To preview, Section 2 elucidates the social cue theory in two main parts. It first explains how social cues work in general terms to lay down the theory’s underlying logic, allowing researchers to apply it to other phenomena. It then develops the theory in the specific context of humanitarian intervention by the community of liberal democracies. Generally stated, the theory interprets legitimization as a social process driven by identity. It argues that institutions, depending on the identities they represent, confer legitimacy by sending social cues about whether a policy is socially appropriate and how a specific group of countries will receive it. These cues operate through relational mechanisms, alleviating concerns about norm abidance, group participation, as well as status and image, and by exerting direct social pressure on group members to conform. Additionally, the theory explains how individuals and communities of ingroup members both can send social cues, but that those cues are more influential when channeled through institutions. When an institution represents a social group, it can amplify and clarify the social meaning of a cue, enhancing its legitimizing effect.

By analyzing institutional legitimization in terms of identity politics, the social cue theory produces a new set of propositions regarding whose cues matter and why. Cues from ingroup members or institutions should matter most because they exert social pressure and induce social considerations regarding a policy. In the case of humanitarian interventions, the relevant ingroup is the community of liberal democracies, which are the main actors behind these interventions, and its closely linked institution, NATO. The theory implies that liberal democracies vis-à-vis NATO can send social cues to influence how domestic and foreign community members perceive humanitarian intervention. This phenomenon occurs because these audiences view NATO as an international organization (IO)Footnote ²⁶ representing their ingroup.Footnote ²⁷ In contrast, cues from outgroup countries and institutions representing a more diffuse community like the Security Council exert less social influence, particularly when an ingroup cue already exists. Section 2 concludes with a discussion of alternative explanations (regarding material considerations and Western regionalism), scope conditions (relating to policy domain, temporality, and political actors), and how the social cue theory is distinct from existing informational arguments.

Sections 3–5 test the social cue theory’s implications through a multi-method research design, using historical polls, case studies of U.S. intervention, news media analysis, a survey of policymakers in the United Kingdom, and nine public opinion surveys and survey experiments conducted in the United States, Japan, and Egypt. The findings from these studies support the social cue theory while refuting alternative explanations.

To begin, during humanitarian interventions, both the Security Council and NATO are highly visible to citizens in liberal democracies – appearing in news articles, speeches, and media broadcasts. As a result, their policy positions become key social cues to the public.

Next, the historical record of U.S. intervention and original survey experiments reveal that ingroup cues significantly influence domestic public opinion. Americans tend to be skeptical of interventions without institutional backing, such as in the case of Rwanda and Syria. They are more enthusiastic about intervention when formal institutional backing is involved but express little distinction between NATO’s sole approval versus both NATO and the Security Council’s backing. Thus, the historical data suggest that ingroup cues from NATO and the liberal community raise domestic public approval, while the Security Council’s additional endorsement has minimal impact. Moreover, survey experiments demonstrate the causality found in the historical record. NATO’s policy endorsements cause an increase in public support for intervention. Furthermore, once an ingroup cue from NATO is sent, the Security Council does little to shift public opinion. The reverse, however, is not true. Even with a Security Council endorsement, NATO’s additional endorsement still increases public approval. Lastly, the experiments also show that social cues from the liberal community influence public opinion, but its cues exert an even stronger influence when conveyed through NATO, demonstrating the power of institutionalized social cues.

Three additional empirical results demonstrate the social cue theory’s causal processes while ruling out alternative explanations relating to military and material burden-sharing capacity. First, NATO’s influence is most substantial among individuals who associate NATO with the liberal community and who express affinity with its member countries, rather than those who associate NATO solely with military strength. Second, NATO’s impact on public opinion operates through the theorized relational mechanisms: concerns about norm abidance, group participation, and status and image. Third, even after removing its ability to change people’s cost-benefit calculations, NATO continues to shape public support for military intervention.

Social cues influence the foreign audiences of an intervening country as well. Analysis of news articles, historical polls, and original survey experiments shows that NATO and the liberal community’s cues similarly affect Japanese support for U.S. humanitarian interventions, just as they do in the U.S. domestic context. These results contribute to debates about American soft power and challenge the notion that NATO’s influence is limited to the West. Turning to foreign elites, a representative survey of UK parliamentarians reveals that they prefer humanitarian intervention with NATO but without the Security Council’s approval over the reverse. In contrast, neither NATO nor the Security Council significantly affects Egyptian public opinion, which instead is more strongly influenced by the Arab League. These findings are consistent with the social cue theory, reinforcing the importance of ingroup cues. Taken together, the evidence shows that the liberal community and NATO can legitimize humanitarian wars by sending powerful social cues to domestic and international audiences.

The final two sections revisit the existing literature on how international institutions affect domestic and foreign opinion on war, demonstrating this Element’s broader contributions to the field of international relations. Section 6 reinterprets existing studies on international law and rational information transmission, showing that evidence previously seen as consistent with legal or material theories can be explained through the lens of social cues. The section also presents additional analysis to test more specific claims made by the existing theories, focusing on political knowledge. These analyses ultimately uncover empirical patterns that can only be explained by the social cue theory. Section 7 discusses the academic and policy implications of the Element. It explains how the Element advances our understanding of legitimacy, forum shopping,Footnote ²⁸ and the role of identity in international relations, and concludes by contemplating practical questions about humanitarian wars led by the democratic community.

2.1 The General Argument

The social cue theory begins with premises rooted in social and political psychology, including social identity theory (SIT), but innovates theoretically to advance a new framework for understanding how institutions convey social cues. People develop social identities when they identify with a particular group. These social groups can form around ascriptive attributes like age, race, and gender as well as ideologies, hobbies, professions, and other qualities. Identification with a group creates a sense of belonging, distinguishing “us” from “them.” This “ingroup-outgroup” distinction shapes people’s understanding of their social world, a concept central to SIT. Scholars of international relations similarly highlight how group identification and ideology foster inclusion, exclusion, and relational comparisons, influencing behavior and preferences in international affairs.Footnote ²⁹

Identifying with a group also means acquiring a stake in the group’s collective well-being and adopting the group’s perspective. For example, individuals may feel pride based on the performance of their sports team, ethnicity, or country. This phenomenon involves developing a sense of social self or collective social identity,Footnote ³⁰ which leads people to experience “collective self-esteem” and emotions based on their group’s experiences.Footnote ³¹ Collective identification also leads people to make intergroup comparisons (which contrasts with comparisons between self and others at the individual level).Footnote ³² They will care about how their group interacts and relates to other groups.

This concept of collective identity is central to international relations scholarship, which shows how the social belonging of countries, governments, and their citizens influence their preferences and behavior in international affairs.Footnote ³³ For example, German citizens adopting their national identity may think about the world from the perspective of their country, Germany, as a social actor in international politics. Research in international relations also emphasizes how social identities can transfer from country to individual: if countries belong to a larger social group, then that country’s citizens will also tend to perceive themselves as belonging to such a group. For example, Germany is a part of Europe, so to some degree, Germans share a collective European identity.Footnote ³⁴ Overall, countries and their citizens can develop ingroup identities based on economic, political, social, and other factors, in addition to geography.

Extensive research documents how people’s identities affect their opinions, decisions, and behavior both as individuals within a group and as someone who has internalized their group’s collective identity. One of the most consistent findings in this area is ingroup favoritism, the tendency to favor those within one’s identity group. This favoritism can be observed in a wide range of actions, from essential services like providing health care to smaller acts of assistance like helping a stranger who has dropped their groceries.Footnote ³⁵ In international relations, this tendency influences humanitarian aid, where individuals are more inclined to help foreigners sharing their racial or religious background.Footnote ³⁶ A notable example arose during the Russian invasion of Ukraine, where a shared European identity contributed to widespread receptivity toward Ukrainian refugeesFootnote ³⁷ – a response that contrasted with the more limited support for Syrian refugees fleeing conflict.Footnote ³⁸ Ultimately, ingroup favoritism is a well-documented phenomenon influencing various aspects of social, economic, and political interactions globally.

Beyond ingroup favoritism, people identifying with a group are more likely to adhere to its views and behave consistently with its norms and practices. This tendency, in part, is driven by self-esteem: people seek to feel good about themselves, and they often adopt their group’s standards as a measure of self-worth and status. For instance, academics are socialized to feel validated by publishing in particular research journals or achieving specific career milestones. From a more rationalist perspective, people’s incentive to affirm their identities can be interpreted as an “identity-based payoff,” where people gain satisfaction from the identity-affirming actions of themselves and others.Footnote ³⁹ Individuals might also follow their group’s behavior and norms for less calculated and more instinctual reasons, out of habit, emulation, and herd mentality. This can include a natural aversion to standing out or appearing deviant within a group to which they feel deeply connected. In this way, both calculated and subconscious motivations contribute to individuals’ alignment with their group’s values and practices.

Following this line of thought, I argue that social cues influence people’s opinions and behavior by operating on their desire to conform to and affirm their identities. The dynamic of social cueing occurs when a “sender” sends a cue to a “recipient,” and the recipient perceives the sender as belonging to its social group. The transmission of a social cue could be direct or indirect. For example, a sender could relay its opinion directly to a recipient, or a sender can publicly declare its opinion on a matter while the recipient observes that declaration. Cues include statements or actions that endorse a point of view, type of behavior, or policy. However, they do not rely on imposing a direct material cost or benefit on a recipient’s behavior, distinguishing them from coercive threats or inducement.

These social cues influence people through two main channels, illustrated in Figure 1. The first channel involves relational mechanisms, which tap into the following three key preferences stemming from people’s social identity. Preference number one is norm abidance, or people’s desire to behave consistently with their group’s norms and practices and advance their group’s values and goals, rather than those of an outgroup.Footnote ⁴⁰ They may find intrinsic value in upholding the group’s values, while they also may fear social disapproval for violating group norms or going against their peers. This adherence can reflect a commitment to a specific norm or a broader, more intangible desire to do what is “right” or “appropriate.” For instance, a Christian may strive to uphold specific norms around charity, marriage, and the family, or more generally, seek to embody what it means to be a “good” Christian. In this sense, social cues push individuals to behave consistently with their group’s values.

Figure 1 Conceptual schema of social cues. Social cues channel their influence through social-relational mechanisms and by exerting direct social pressures to conform.

Preference number two is group participation, or people’s desire to be part of their group’s activities. People often prefer to do things alongside their peers, whether going out for a movie or engaging in a multilateral foreign policy. On the other hand, people may also have a “fear of missing out” when they are left out of the activities of their social group. Social cues thus induce this desire to be part of the community.

Preference number three is status and image, or people’s desire to maintain a high status and a good image among their peers. For many, maintaining a good image serves not only as an intrinsic source of self-esteem but also as an instrumental way to secure a sense of high status within a group.Footnote ⁴¹ This desire can lead to social ladder-climbing behaviors as well, as individuals seek recognition and approval. Furthermore, this drive for status and a positive image often extends beyond the individual to the group as a whole: people want their group to be respected within an even larger, superordinate community. For instance, a tennis player may want to be well-regarded among her fellow players while desiring that tennis itself to be well-regarded relative to other sports. Thus, social cues act upon these concerns about status and image, which in turn influence people’s opinions and behaviors. Together, these three mechanisms – norm abidance, group participation, and status/image – help to explain how social cues influence people’s opinions and behaviors.Footnote ⁴²

The second channel of social cues operates more directly, affecting individuals through direct social pressure. In this case, social cues prompt recipients to adopt certain beliefs or behaviors without explicit reward, punishment, or incentive. Instead, recipients change their opinions or behavior through an almost automatic or instinctual process of emulation, mimicking, and conformity. This direct pressure might be described as a “normative nudge,” and there is no intervening mechanism in its causal pathway. In summary, social cues play a dual role of social influence: they ameliorate social-relational concerns, while also directly socializing group members into conformity.Footnote ⁴³ From the surface, this dual role can be interpreted as conferring legitimacy on a point of view or behavior among a particular social group.

Who or what types of entities can send social cues? The sender of these cues can be individuals, groups of individuals, or formal bodies or organizations. Among these, cues from ingroup members exert the strongest influence. Indeed, experimental research shows that people update their beliefs and change their behavior after gleaning information from fellow group members.Footnote ⁴⁴ Other studies show that cues from the like-minded can influence people’s trade attitudes.Footnote ⁴⁵ Yet other research finds that cues from social peers can influence public opinion as strongly as those from political elites.Footnote ⁴⁶ Within a group, cues from veteran or high-status members tend to be more influential than those from novices.Footnote ⁴⁷ By contrast, cues from outgroup members are generally discounted or even met with resistance, especially when the outgroup is perceived negatively or is socially distant.Footnote ⁴⁸

Social cues are particularly impactful when they are institutionalized. In international relations, for example, international organizations (IOs) can institutionalize the cues sent by certain countries or groups of countries. In this context, institutionalization designates a cue-sender as representing a particular social group using rules, titles, or formal organizations. Furthermore, such a designation can elevate the status of a cue-sender within their community. For example, a religious leader with a specific title can send cues about the social implications of various behaviors for members of a particular religion. The role of a religious leader is institutionalized (e.g., a church pastor or an abbot of a temple), which places them in a better position than other ingroup members to send social cues about what constitutes identity-congruent behavior. To provide another example, the European Union and its high commissioner are formal institutions that could effectively send social cues to the community of European states and citizens, compared to other ingroup members like the German government or some high official that happens to be European.

Institutions, when recognized as representing a social group, can amplify the impact of a social cue in two related ways. First, institutionalization clarifies the social meaning of a cue. Individuals, countries, and other entities often have multiple identities, so a cue’s social meaning may be ambiguous to observers. Take, for example, a pastor named Mike, who is also a parent, political party member, and engineer. If Mike, dressed in everyday clothes, shares a political opinion at a town hall meeting, fellow religious group members might be unsure whether to interpret his views through a religious lens or as influenced by his other roles. However, if he expresses the same views from the pulpit, members of his religious community are likely to interpret and respond to his statement within the context of their shared faith and identity.

Second, institutions can play a “logistical” role by coordinating group decision-making and facilitating the delivery of a social cue to group members. Even if individuals (or states) share an identity, they may have difficulty agreeing on a group’s goals, agenda, and actions. And even if they are aligned, they may struggle to effectively communicate their views to the broader group, especially if the group is large or dispersed. Institutions address these challenges by facilitating group decision-making and promoting communication through channels like traditional and internet media, mailing lists, and gatherings, enhancing the visibility and reach of its viewpoints. So, overall, social cues sent through institutions representing an identity group are likely to have greater reach and provide clearer guidance on ingroup norms.

To summarize, the social cue theory argues that people are more likely to support policies and engage in behavior endorsed by their social group while discounting the opinions of outgroup members. These endorsements, or social cues, influence ingroup behavior both directly – through a “normative nudge” effect that encourages emulation and conformity – and indirectly by mitigating relational concerns about norm abidance, group participation, and status and image. Social cues are particularly strong when delivered through institutionalized channels, such as formal organizations representing the group. This process of social cueing ultimately serves to legitimize certain behaviors and viewpoints for group members.

Before turning to the issue of humanitarian intervention, it is crucial to elaborate on the social cue theory’s main requirements. For social cues to work, cue recipients need to identify with the social group of the cue sender, but only to a minimal degree. Namely, cue recipients need only have some sense, whether explicitly or subconsciously, that they belong to the group. This standard most closely aligns with SIT, particularly from the minimal group paradigm tradition.Footnote ⁴⁹ From this perspective, even trivial things like being sorted into the blue or red group can be the basis of group identity,Footnote ⁵⁰ though if the identity marker is too trivial, it may not be a stable and long-lasting identity. Similarly, for institutionalized social cues to work, ingroup members need to have some sense, whether explicitly or subconsciously, that the institution is associated with their social group.

Two important points follow. First, members of a social group can still be susceptible to social cues even if they lack detailed knowledge of their group’s norms. For example, a person may consider themselves a member of the U.S. Republican Party. Similarly, someone who identifies as Christian may not be familiar with all the ways their beliefs and actions align (or don’t) with their faith. However, both of these individuals would still be susceptible to social cues by their group. Second, a social group can form around a certain set of values but abandon those values over time while still waving the group’s banner. This phenomenon could lead to organized hypocrisy in identity politics. Thus, one can identify with a social group and be susceptible to its social cues without necessarily understanding, internalizing, or keeping pace with its evolving values and practices.

2.2 Social Cues by the Liberal Community and NATO

In three parts, I will now apply the general formulation of the social cue theory to international politics and, specifically, to understanding how international institutions legitimize and, therefore, influence the domestic politics of humanitarian wars waged by liberal democracies. This discussion will generate a series of hypotheses regarding how the Security Council, NATO, and their member governments affect people’s views on humanitarian intervention, which will be tested in the subsequent empirical sections.

The first part of the argument establishes the liberal democratic community and NATO as the social group and institution most central to the phenomenon of armed humanitarian interventions. The United States, along with other liberal democracies, is the primary participant of humanitarian interventions, and substantial research on political communities suggests that these countries are embedded in a broader social grouping of democratic countries. This democratic community distinguishes itself from outgroup countries governed by authoritarian regimes, closed economic systems, and those with limited fundamental rights.Footnote ⁵¹ This us versus them perspective was undoubtedly promoted during the Cold War, with both sides framing the world as ideologically divided between good and bad systems.Footnote ⁵² While such a perspective may have been elite-driven and motivated by material interest, the notion that democracies share an identity gained traction and continued to shape foreign policy and identity even after the fall of the Berlin Wall.Footnote ⁵³ This democratic identity also influences how various political actors, including the public, perceive which states are “friends” or “foes.”Footnote ⁵⁴ Indeed, reflecting on the resilience of political identities, Peter Katzenstein noted that political actors “attribute far deeper meanings to the historical battles that define collective identities than to the transient conflicts of daily politics.”Footnote ⁵⁵

Beyond representing a distinct social group, the liberal community is defined by a shared set of norms and practices.Footnote ⁵⁶ Members follow a “logic of appropriateness,” or a set of beliefs about how they should behave,Footnote ⁵⁷ which may be formally codified or informally accepted and adhered to as part of a “community of practice.”Footnote ⁵⁸ For liberal democracies, this means governments and their citizens alike tend to avoid coercive bargaining or violence in interstate relations, although this courtesy may not extend to interactions with outgroup nations.Footnote ⁵⁹ Relevant to this study, these democracies also uphold norms of consultation, emphasizing the need for collective deliberation with other democratic states in foreign policy.Footnote ⁶⁰ As a result, democracies often take the policy endorsements of other group members into serious consideration.Footnote ⁶¹

Other research demonstrates how the social distinction between democracies and non-democracies influences international politics. On a macro level, democratic ideological and normative group distinctions affect which states wage war with one another,Footnote ⁶² form military alliances,Footnote ⁶³ and perceive other governments as threats.Footnote ⁶⁴ An analysis of UN General Assembly voting records even reveals that commitments to liberalism create a coherent “liberal order” grouping in international affairs.Footnote ⁶⁵

Group behavior among countries – whether it be the deliberation of policy or conduct of joint operations – often occurs within a multilateral IO. When it comes to the democratic community, a long tradition of scholarship observes that NATO is emblematic of this group of countries, particularly in contrast to other IOs like the Security Council.Footnote ⁶⁶ This line of research, often traced back to Karl Deutsch and colleagues’ seminal 1957 study of the North Atlantic community, links NATO’s role its member countries’ commitment to shared norms and values.Footnote ⁶⁷ Subsequent research shows how NATO’s function as a political community of norms and values further sheds light on its operational practices, institutional survival,Footnote ⁶⁸ as well as how its members maintain peace and engage in conflict.Footnote ⁶⁹ NATO’s expansion into Eastern Europe can be understood in terms of social identity community building by liberal democracies as well.Footnote ⁷⁰ Thus, while the democratic community exists as a group of countries, NATO has become a key institution embodying its social identity and place in international relations.

To be clear, understanding NATO as a social community does not negate its role as a military alliance. These functions are not mutually exclusive and may even be complementary. In fact, even NATO itself recognizes its dual mission of providing security and advancing democratic norms and group cohesion. Since the end of the Cold War, NATO has explicitly placed promoting and defending liberal democratic values and community on par with its security mandate. This is evident in NATO’s Strategic Concept documents, which serve as its primary public statements that articulate its purpose, principles, and goals. The early Strategic Concepts (1950–1968) emphasized NATO’s strength as a military alliance, though they still acknowledged its interest in safeguarding democracy. Beginning with the fifth Strategic Concept in 1990, however, NATO has increasingly emphasized its social role in building and embodying a “shared community” of democratic values. The 2022 Strategic Concept, for example, states on its first page:

This statement underscores NATO’s dual identity, reflecting its mission to uphold both democratic ideals and military security.Footnote ⁷¹

In sum, NATO can be understood as an ingroup IO for democratic nations.Footnote ⁷² In contrast, the Security Council has a more diffuse and less defined social identity, which, according to social cue theory, limits the influence of its cues. While the Security Council includes some democratic members, it also represents a range of nondemocratic states and thus reflects a broader range of the international community. Indeed, a defining feature of the Security Council is its heterogeneous membership.Footnote ⁷³ If it exclusively represented nondemocratic governments, it might convey an outgroup social cue, but as it stands, it occupies an ambiguous position in terms of ideological and social alignment with the democratic community.

The second part of the argument turns to the micro-level and assesses how the democratic community and NATO’s representation of it manifests within individuals. For the social cues to influence attitudes toward humanitarian interventions, a meaningful segment of people within democratic countries must identify, at least loosely, with the broader liberal community. According to the social cue theory, individuals need only a minimal or “thin” identification with a group for such cues to take effect. Often, people associate with a group before internalizing all the group’s values. This means that not all members of the liberal community will necessarily understand or strictly adhere to its norms, practices, and ideals. Rather, it is enough that people feel, even if just implicitly, that other liberal democracies and NATO are “on the same team” and constitute their ingroup. By holding this basic degree of identification, individuals become susceptible to social cues, though a stronger group identification may amplify these effects. This section supports the plausibility of these assumptions regarding people’s identity attachments.

A substantial body of research theorizes and documents how identities in international relations manifest at the individual and mass public levels.Footnote ⁷⁴ Within the liberal community, people’s values and identification with other democracies are shaped by elite communication, news media, education, and public discourse.Footnote ⁷⁵ Studies show that democracy and its values affect both mass public and policymaker attitudes toward international affairs.Footnote ⁷⁶ For example, research from the transpacific region shows that democratic leaders’ appeals to shared liberal values affect citizen attitudes toward military alliances.Footnote ⁷⁷ Relatedly, global opinions about China’s rise as a superpower vary significantly based on people’s democratic orientation, even when controlling for national security and economic interests.Footnote ⁷⁸

Corroborating these findings, Figure 2 presents data from a survey of Americans and shows that about 78 percent of the respondents believed that sharing a political system was at least “Somewhat Important” in determining whether citizens of different countries could relate to one another. Eighty percent saw language, a key marker of shared identity, as important.Footnote ⁷⁹ In contrast, only about 45 percent considered race an important basis for international relations.

Figure 2 Americans can better relate to foreigners from democratic systems. Question asked, “What allows people from different countries to relate with one another?” The percentages capture those who responded somewhat, very, or extremely important, as opposed to not very or not at all important. N = 704. Survey USA-5.

Additional research documents how NATO membership affects public beliefs and values related to liberal democracy. For example, after joining NATO, countries like the Czech Republic and Romania adopted textbooks and implemented educational campaigns that promoted democratic frames of thinking, prompting citizens and elites to prioritize and see liberal values and human rights as legitimate.Footnote ⁸⁰ This experience reflects a broader pattern of liberal democratic community-building in Eastern Europe, where NATO has served as a socializing force, fostering identification with democratic norms and values.Footnote ⁸¹

More recent examples from NATO’s public diplomacy provide additional evidence of how it fosters a collective identity among individuals. NATO states, as a core tenet, that it “promotes democratic values and enables members to consult and cooperate on defense and security-related issues.”Footnote ⁸² To this end, NATO and its affiliates often portray the alliance as a value-driven, group-oriented community. For instance, in June 2017, NATO tweeted, “We are an Alliance of like-minded countries … We are united … #WeAreNATO.”Footnote ⁸³ Another tweet from the U.S. Mission to NATO highlights the group’s ideational commitments, “SecDef Mattis: For nearly 70 years the #NATO alliance has served to uphold the values upon which our democracies are founded.”Footnote ⁸⁴ Similarly, U.S. Ambassador to the United Nations, Linda Thomas-Greenfield, tweeted, “NATO is the most powerful and successful alliance in history, and it’s built on the foundation of shared democratic values.” Beyond these social dynamics, NATO’s attempt to foster a collective identity can also be understood in strategic terms: if its members share an identity, they are more likely to contribute to the organization’s overarching goals.Footnote ⁸⁵

Whatever the motivation, these acts of socialization lead individuals in the community to identify with the larger group. Indeed, American public opinion reflects NATO’s dual role as a military alliance and emblem of the democratic community. Survey results summarized in Figure 3 show that most Americans associate “NATO” with the terms friends, democracy, or military. Notably, they are more likely to associate NATO with ideas relevant to a liberal community, such as friends and democracy, than they are with the Security Council. Another telling result is that friends is the most frequently selected term, aligning with the concept of ingroup identification and reflecting NATO’s role as more than just a defense alliance.

Figure 3 Americans associate “NATO” with the military, democracy, and friends. Question asked, “What do you associate with the [North Atlantic Treaty Organization (NATO)/United Nations Security Council]?” Response options were randomized. N = 1,790. Data are from Survey USA-6.

The third and final part of the argument explains how the liberal democratic community sends social cues regarding humanitarian intervention to ingroup members. Here, countries within the liberal community – especially those institutionalized through NATO – act as cue senders, while citizens in these countries, from laypersons to policymakers and elites, are the recipients. In this setting, social cues take the form of policy positions adopted by foreign governments and institutions like NATO.

Social cues are observed in legislation, legislative votes, and statements or speeches by state officials and leaders, among other behaviors. More broadly, governments and international institutions convey their perspectives through political communication and news media, influencing public discourse. For instance, NATO sent a social cue when the North Atlantic Council voted to authorize intervention in Kosovo in 1999, a decision relayed through news media and political speeches, such as President Bill Clinton’s address to the American public. By contrast, an example of a social cue sent through legislative voting includes the UK Parliament’s 2013 vote against a U.S.-led intervention in Syria, which was covered in American news media.Footnote ⁸⁶ Even outside institutional frameworks, individual foreign leaders and other elites can also directly influence public opinion, as Hayes and Guardino (Reference Hayes and Guardino2013) document in their study of the 2003 Iraq War. These policy positions can then be channeled to citizens through the media and the rhetoric of domestic elites attempting to bolster policy support by referencing the endorsements of “foreign voices.”

The liberal community’s policy endorsements act as social cues that influence members of its ingroup. These social cues exert social pressure, prompting group members to adopt the sender’s stance on humanitarian intervention. As theorized, social pressure directly prompts an instinctual alignment with a particular set of preferences or behaviors. They also indirectly mitigate social-relational concerns about the implications of intervention for members of the liberal community. In doing so, social cues address three key questions about how intervening would affect one’s country. Is intervention in line with group norms, which include championing liberal values and human rights, and is it simply appropriate or the right thing to do? Will other ingroup countries also participate in the intervention? Will engaging in intervention harm or improve their country’s status or image?Footnote ⁸⁷ The answer to these questions influences people’s understanding of pro-norm behavior, group belonging, and status and image, ultimately influencing their support for humanitarian intervention policy.

Next, we can elaborate on the specific role of NATO in institutionalized social cues by the liberal community. As discussed in the general theory, NATO can clarify the social meaning of cues sent by its member states while also amplifying the reach of these policy endorsements. First, in clarifying social meaning, NATO can help audiences interpret the motives behind a member state’s actions. When a state endorses a course of action, it may not be immediately clear whether the support reflects the broader liberal democratic values or narrower national interests. For instance, if the UK backs military intervention, it could be for specific national interests rather than a stance on behalf of the democratic community. However, when the UK makes the same recommendation under NATO’s banner, the endorsement takes on a more communal significance, as NATO symbolizes the collective liberal democratic community. In this way, NATO strengthens the social cue by clarifying its social meaning within the democratic community.

Second, NATO serves an amplification role, enhancing the logistical coordination and communication reach among member states and the public. For governments, NATO helps coordinate decisions across the community,Footnote ⁸⁸ increasing the likelihood of reaching shared policy positions, such as on intervention issues. Without this organizational framework, individual governments may struggle to reach or communicate collective decisions. For the public, NATO can increase the prominence of the democratic community’s policy cue, as its media presence often surpasses that of any single member state.Footnote ⁸⁹ The following empirical sections highlight how NATO receives substantial coverage in both print and television media, far outpacing individual member states. This visibility supports NATO’s role in amplifying the community’s voice, ensuring that its policy endorsements reach a broad audience.

In conclusion, social cue theory offers a coherent and distinct perspective on the social dynamics within the liberal community regarding humanitarian interventions. It suggests that the liberal community can influence democratic citizens through ingroup social cues, which gain even greater power when channeled through institutions like NATO. Furthermore, once citizens receive such a policy endorsement (or rejection) from NATO, they are more likely to discount the views of outgroup members or institutions with mixed identities, such as the Security Council. These cues are influential because they exert social pressure and provide social guidance, helping individuals understand how to fit in and gain good standing within the community of democracies.

2.3 Caveats and Clarifications

This section discusses alternative explanations for NATO’s apparent legitimizing effect, the generalizability of the theory, and how the social cue theory contrasts with existing informational theories of international institutions.

2.4 Outline of Empirical Sections and Research Designs

The subsequent empirical sections test various implications of the social cue theory as outlined next. Section 3 assesses the effect of social cues sent by the liberal community and NATO in the U.S. context. It does so by examining historical cases and original experimental data from surveys. Section 4 demonstrates the mechanisms of social cues, distinguishing the social causal pathways from material alternative explanations. Section 5 examines the effect of social cues on foreign audiences, and Section 6 (re)assesses arguments from the extant literature relating to international law and information transmission.

• Observational evidence from American interventions (Section 3): Historical polls from 1990 to 2013 show that humanitarian interventions backed by the liberal community and NATO enjoy high support from the American people. The Syria case shows that IO approval is critical for raising public support. Comparing Bosnia and Kosovo shows that NATO, with or without the Security Council, can generate public support.

• Causal evidence of social cues on domestic audiences (Section 3): Survey experiments show that cues from NATO and the liberal community cause American support for humanitarian intervention to increase. Once Americans learn about NATO’s position, the Security Council has little effect, but not the reverse. These effects are even stronger among the attentive public than among the mass public, implying that NATO’s cues are not primarily serving as informational heuristics for the ignorant, as existing theories might imply.

• Institutionalized cues (Section 3): Additional experiments show that the liberal community’s influence on public opinion is more substantial when institutionalized via NATO. Corroborating these experimental results, during the Libya crisis, NATO also received more cable news television coverage than its component countries, showing that IOs receive more salience than specific countries in the real world.

• Analysis showing the relevance of ingroup perceptions to NATO’s influence (Section 4): NATO’s endorsement effect is greatest among (1) those who express the strongest affinity with NATO’s member countries and (2) those who associate NATO with democracy and community. In contrast, NATO’s influence on public opinion is not impacted by its association with military power, contrary to the material considerations alternative explanation.

• Analysis of the social cue’s relational mechanisms (Section 4). NATO’s cues activate people’s social concerns about norm-abidance, group participation, and status and image. These factors, in turn, affect people’s opinions on intervention. In contrast, material considerations relating to financial and human costs and benefits cannot fully explain the cueing effect.

• Foreign audiences (Section 5): The relative effects of the liberal community, NATO, and the Security Council, as well as the impact of institutionalized cues, reported in Section 2, are all replicated in Japan. This finding contrasts with the alternative explanation of Western regionalism. Next, NATO and the Security Council do not significantly affect Egyptian public opinion, but the Arab League does. This finding is also consistent with the social cue theory.

• Foreign elites (Section 5): Substantially more members of the United Kingdom Parliament (MPs) would rather receive NATO’s backing for humanitarian intervention than the Security Council’s if they could only have one or the other.

• Reconsidering legal perspectives (Section 6). This section provides additional observational and experimental evidence showing that the Security Council’s lack of influence is not due to people’s ignorance of international law.

• Reconsidering arguments about information transmission (Section 6). Existing theories predict that a neutral and conservative institution like the Security Council should affect public opinion, while a relatively homogenous and hawkish institution like NATO should have a weaker effect. Yet, the evidence thus far shows nearly the opposite. This section presents additional analysis to show that the Security Council’s relatively weak effect is not due to political ignorance. People largely view the Security Council as relatively neutral and conservative. Individual-level data on people’s understanding of the Security Council and NATO further show that people’s assumptions about institutional neutrality and conservativeness do not moderate the institution’s effect on public opinion.

3.1 The Post–Cold War Historical Record

Since the end of the Cold War, armed humanitarian interventions have primarily been a policy of the liberal democratic community, with the United States leading the majority of these actions. Each of these interventions has been multilateral, conducted under the auspices of an international organization such as the Security Council or NATO. Conversely, intervention efforts have been abandoned when institutional approval proved unattainable. This section reviews historical opinion polls from episodes of U.S. interventions, focusing on the case of Syria and a comparative analysis of Bosnia and Kosovo. Syria highlights the challenges of rallying support for unilateral intervention, while the Balkan cases illustrate how NATO can drive humanitarian action with or without Security Council endorsement.

What do the broad, descriptive patterns tell us about the post–Cold War era of humanitarian intervention? Table 1 summarizes the historical relationship between the Security Council, NATO, the liberal community (independent from NATO), and American public support for humanitarian intervention.Footnote ¹⁰⁹ It includes cases in which the United States considered humanitarian intervention and opinion polling data were available. Column 1 names the case. Columns 2 through 4 give the policy position of the Security Council, NATO, and the liberal community. NATO’s policy position on intervention is generally equivalent to the policy position of the liberal community. But when NATO did not consider the case of intervention (i.e., when NATO is N/A in the table), the country positions of the liberal community are coded using public statements and actions of NATO’s member states. Finally, Columns 5 and 6 give the percentage of respondents who supported intervention, both for each case and for the aggregation of cases that fall under the same category of multilateralism. The Online Appendix provides detailed notes for each case, its coding, and survey wording.

The record reveals three broad categories of humanitarian intervention: those with no systematic international support (e.g., Syria), those backed by NATO or the club of democracies but not the broader international community via the Security Council (e.g., Kosovo), and those with widespread international backing (e.g., Bosnia). Public opinion across these categories suggests a clear preference among Americans for interventions with international backing. Only 28–31 percent of the public supported intervention without foreign approval, while a majority supported interventions with some degree of IO approval (about a 22–25 percentage point difference in support). Interestingly, in comparing the second and third categories of interventions, Americans do not appear to differentiate interventions with both Security Council and NATO approval from interventions with only NATO approval.Footnote ¹¹⁰ These patterns are consistent with the identity theory’s prediction that once a cue from the ingroup is received, the additional cue from other countries contributes minimally to public support.

3.2 Experimental Evidence

The historical record suggests a significant role of institutional legitimacy – particularly the approval of fellow democracies and NATO – in shaping public support for humanitarian intervention. Nonetheless, questions remain about whether international institutions cause public opinion to change. First, one may wonder if public opinion itself influenced institutional decisions, rather than institutions shaping public opinion. For instance, in Syria, perhaps France and Germany hesitated to support intervention because they perceived that American public reticence would be challenging to overcome, even with their endorsement. Second, others may point to idiosyncratic factors rather than systematic patterns about NATO and the Security Council as the primary drivers of public opinion. The scarcity of cases could make it particularly difficult to refute this critique. For example, Somalia was the U.S. first major humanitarian intervention following the widely popular Gulf War, and the public’s enthusiasm may have been fueled by the “warm glow” of that recent victory, rather than by any international consensus. In the case of Syria, one might argue that religious bias, rather than the lack of institutional backing, reduced public support for intervening on behalf of the primarily Muslim victims of civil war.Footnote ¹³¹ Both selection bias and case-specific idiosyncrasies could, therefore, interfere with establishing causality.

To address these questions, the following section turns to experimental evidence from original surveys to assess whether NATO’s cues directly influence public opinion. In addition, these survey experiments then disentangle the effect of the broader community of democratic countries from their institutionalized cues sent through NATO – something historical data alone cannot achieve.

3.2.2 NATO and the Security Council’s Cueing Effect

The first survey, administered to a nationally representative sample of Americans by YouGov, allows us to estimate the main effect of IOs on public opinion (Survey USA-1). Figure 4 summarizes its results and shows that moving from a situation where NATO and the Security Council oppose intervention (Both Oppose) to one where NATO endorses intervention but not the Security Council (NATO Only) raises Support by 26.4 percentage points (pp.). This difference is a substantial and statistically significant effect (p<0.01). By contrast, the additional impact of the Security Council’s endorsement (moving from NATO Only to Both Endorse) is substantively small (4.3 pp.) and statistically insignificant (p = 0.20). Thus, as hypothesized, an ingroup cue from NATO raises public support for war, and once the ingroup cue is received, the broader cue from the Security Council has little effect.

Figure 4 NATO raises support for intervention, the Security Council adds a little extra. This figure shows the percentage of respondents supporting intervention depending on NATO and the Security Council’s stance on intervention. 95% confidence intervals are given. N = 1,000. Data are from Survey USA-1.

Bolstering the external validity of these findings, the 26.4 pp. effect derived from the experiment mirrors the historical polling data shown in the previous section, which estimated a 22 pp. difference between the NATO only and dual endorsement scenarios. Furthermore, these effects are robust to (if not even larger) examining only the informed public or liberal internationalists, implying that the effect is present among those informed about international affairs.Footnote ¹³² NATO’s more substantial impact on the informed public also implies that its cues are not exclusively operating as a heuristic to solve the problem of rational ignorance, as some informational theories would predict.

I now turn to Survey USA-2, which replicates the three scenarios in the first survey but adds a fourth scenario in which only the Security Council but not NATO supports intervention (UNSC Only). This scenario is unusual from a historical standpoint since there is no case in which NATO rejects intervention but the Security Council, which includes veto-holding NATO members, endorses intervention. Nevertheless, this hypothetical situation can be used to evaluate the social cue theory. Specifically, critics may argue that the results displayed in Figure 3 should be interpreted as a generic first versus second cue effect. The fact that receiving NATO Only has a large effect and IOs Endorse does not further boost public opinion is an artifact of NATO providing the first cue, while any second cue will be less valuable. While plausible, the data do not support this critique.

The left-hand side of Figure 5 shows the level of public support in each of the four conditions from Survey USA-2. It replicates the main results from Survey #1, where NATO Only was substantially higher than Both Oppose, but Both Endorse is trivially greater than NATO Only. In contrast, while UNSC Only increases public support, the additional endorsement of NATO in Both Endorse still generates additional (and even more) public support. The right-hand side of the figure combines NATO and the Security Council’s effect sizes conditional on each other into an average effect. It shows that the effect of NATO is 15.7 pp., and the Security Council is 6 pp. These results demonstrate that NATO exerts greater influence on American public opinion than the Security Council, not just when it is the first policy endorser. Furthermore, these data refute the view that citizens cannot distinguish between NATO and the Security Council. Finally, these results cannot be easily explained by legal and informational theories, which hypothesize that an affirmative endorsement by the Security Council should raise public support for a policy. The results are instead consistent with the social cue perspective on the centrality of ingroup cues.

Figure 5 NATO has a greater effect than the Security Council. The figure on the left shows public support for intervention in four scenarios regarding NATO and the Security Council’s stance on intervention. The figure on the right reports the average treatment effect of each IO. 95% confidence intervals are given. N = 408. Data are from Survey USA-2.

3.2.3 The Power of Institutionalized Cues

In the first two surveys, survey takers read information about the IOs and their member countries in conjunction. This research design choice increases the realism of the survey: news and political speeches relating to IOs and military intervention generally mention both the IO and key countries.Footnote ¹³³ However, such a design cannot disentangle the effect of the IOs from their component countries, which is required to evaluate the social cue theory’s argument about the power of institutionalized cues. To answer these questions, this section thus turns to additional experimental and observational analysis.

Beginning with the experimental data, a third survey seeks to disaggregate the effect of NATO’s cue from the effect of its member countries sending a cue without NATO (Survey USA-3). The survey experimentally split the respondents into three groups, each with two sub-conditions (3x2). As in the previous surveys, the first one-third of the respondents were experimentally assigned to read that “The North Atlantic Treaty Organization (NATO) [supports/opposes] taking military action to help these civilians. NATO members include the U.S., Canada, and several European countries.” Then, to distinguish the effect of NATO from the effect of its named countries, the second one-third of the respondents read about the policy position of NATO without mention of any particular country, while the final one-third of the respondents read about the policy position of “U.S., Canada, and several European countries” without mention of NATO.

Thus, this design creates six experimental groups: Only Countries, Only NATO, or Both Countries and NATO either support or oppose military action.Footnote ¹³⁴ As before, the dependent variable measures public support for intervention. Figure 6 summarizes the results of this survey. When the liberal community sends a social cue (Only Countries Named Support – Oppose), public support for war shifts by about 15.5 pp. When it sends the same cue via NATO (Both Named), public support shifts by about 32.8 pp. The effect of naming only NATO is included to be comprehensive but is challenging to interpret. Even though it does not mention NATO’s member countries, respondents can easily infer that several democratic countries support intervention simply from hearing “NATO.” What this data does help to clarify, though, is whether the cue’s effect relies primarily on naming countries. The results show they do not: the cue effect under that condition is 24.8 pp., which is less than 32.8.

Figure 6 The liberal community’s cue has a stronger effect when it is sent through NATO. This figure shows the cue effect on public support for intervention in percentage points, depending on whether countries in the liberal community, NATO, or both are explicitly named in the experiment. N = 598. Data are from Survey USA-3.

Lastly, the social cue theory states that institutions do not just clarify the social meaning of a cue, but they can also play a logistical and communication role by increasing the cue’s reach. Without NATO, individual governments in the liberal community might not otherwise be able to coordinate on a policy position and subsequently ensure that their policy viewpoints reach the ears of various domestic audiences. This point cannot be tested in a survey experiment, since all survey takers are shown the social cue directly. Furthermore, it is difficult to establish the causal effect of institutionalization on a cue’s accessibility and reach in the real world, as it is impossible to randomize the existence of NATO. Nevertheless, observational data can help to assess the claim’s plausibility. Figure 7 summarizes data from American cable television collected by the Stanford Cable TV News Analyzer. The data capture the number of minutes per day certain topics are discussed on CNN, MSNBC, and Fox News. According to this data, during the 2011 Libyan intervention episode, NATO generally received more airtime on American cable TV than the individual NATO countries advocating for intervention, France, and the United Kingdom. This pattern is consistent with the theory’s claim that institutionalized cues receive more exposure among the general public.

Figure 7 NATO receives more cable news airtime than its member countries. This figure measures the TV news salience of NATO versus specific NATO countries surrounding the Libyan intervention episode. The graph is produced by the Stanford Cable TV News Analyzer, available at (accessed on 10 November 2024): https://tvnews.stanford.edu/. The web tool is for “count[ing] the screen time of who and what is in cable TV news … The dataset includes near 24-7 recordings of CNN, Fox News, and MSNBC.”

Taken together, the historical and experimental evidence in this section demonstrate that the liberal community and NATO have a powerful effect on American attitudes toward humanitarian intervention, one that is potentially even stronger than the Security Council. Furthermore, cues by the liberal community have more impact and greater reach when channeled through NATO.

4.1 Test #1: Examine Subgroups That View NATO as an Ingroup

The social cue theory implies that NATO’s effect is most substantial among Americans who identify with its member countries and view NATO as an ingroup community of fellow democracies. I test this hypothesis in two ways. First, I examine how NATO’s effect on public support for military intervention varies depending on individuals’ affinity with NATO member countries. To do so, in Survey USA-2, respondents were asked whether they perceive Canada, Germany, France, and the United Kingdom as having a “friendly,” “neutral,” or “hostile” relationship with the United States.Footnote ¹³⁵ For the analysis, I then divided the sample into two groups based on respondents’ average affinity with these four countries.

Figure 8 shows that NATO’s cueing effect is about 6.9 percentage points (pp.) among those with relatively low affinity for the four NATO countries, whereas it reaches 22.6 pp. among those with high affinity for these countries. The difference in effect sizes is substantially large, about 16 pp. (p = 0.086). These findings support the theory: NATO’s cues have a notably stronger impact on Americans who feel a closer connection to its member countries.

Figure 8 NATO’s effect is greater among Americans expressing high affinity with NATO’s members. This figure reports the effect of NATO’s endorsement on public support for intervention among survey takers with Low or High Affinity with NATO’s member countries. 95% confidence intervals are given. N = 408. Data are from Survey USA-2.

Turning to a second approach, I estimated the effect of NATO’s cue on different groups of people, depending on how they understood NATO (Survey USA-3). To measure people’s understanding of NATO, the survey asked, “When you hear ‘NATO,’ which of the following words or phrases come to mind? Please select all apply.” The survey taker could then choose from a list of choices that include friends and democracy, which were relevant to the liberal community concept, in addition to military, agriculture, foreign investment, freeloaders, enemies, none/don’t know. The order of these choices was randomly displayed. About 80 percent of the respondents selected friends, democracy, or military, consistent with the “correct” understanding of NATO.Footnote ¹³⁶

With this data, I classified individuals who selected “friends” or “democracy” as explicitly perceiving NATO as a liberal community. I then estimated NATO’s cue effect among this group. The social cue theory predicts that NATO will have a more substantial impact on this group than those outside of it. To examine the material considerations alternative explanation, I also classified individuals who selected “military” as seeing NATO from a security perspective and then similarly estimated NATO’s impact on members of this group compared to those outside of it. Figure 9 displays the results of this analysis.

Figure 9 NATO’s effect is stronger among Americans who associate it with community, not military. This figure reports the effect of NATO’s endorsement on public support for war depending on whether or not respondents associated NATO with “democracy” or “friend” (i.e., community) or “military.” 95% confidence intervals are given. N = 598. Data are from Survey USA-3.

Consistent with the social cue theory, NATO’s effect is about 15 pp. among those who did not associate NATO with the words friend or democracy, while it is nearly double, about 30 pp., among those who do. Thus, NATO’s impact on public opinion operates not only via general affinity, as the previous analysis in Figure 8 showed, but specifically via ingroup belonging over the substantive issue of democracy.

By contrast, whether people associate NATO with the military or not, NATO’s cueing effect is about the same. If anything, associating NATO with the military reduces NATO’s influence. This dampening effect is about −6 pp., though it is statistically indistinguishable from zero (p = 0.42). These patterns show that NATO’s cueing effect cannot be explained by Americans liking the idea of a strong military organization backing their foreign policy. On the contrary, Americans may possibly be sensitive to the prospects of their country warmongering abroad, so NATO’s military reputation can be detrimental to its influence on domestic legitimacy. While the social cue theory needed not to be incompatible with the alternative explanation of military capabilities, as both logics could operate simultaneously, these findings ultimately only confirmed the social perspective.

4.2 Test #2: Directly Estimate Social Causal Mechanisms

This second test analyzes six causal mechanisms (or mediators) that might explain NATO’s effect on public support for war.Footnote ¹³⁷ The first three relational mechanisms are associated with the social cue argument. The mechanism Norm Abidance captures beliefs about whether intervention conforms to liberal humanitarian norms about using military force for humanitarian reasons.Footnote ¹³⁸ Group Participation indicates respondents’ beliefs about whether other countries will join the operation. Benefit Status/Image measures perceptions about whether an intervention would improve or damage the U.S. “reputation.”Footnote ¹³⁹ The next three mechanisms are unrelated to the social cue theory, though the social cue theory does not rule them out either. The mechanism Prevent Contagion indicates people’s expectations about whether not intervening would lead to a spread of conflict. Retaliation Unlikely records beliefs about whether other countries would punish the United States for using military force. Casualties Unlikely measures expectations about American casualties.

To what extent does each of these six causal mechanisms serve as a link between NATO’s cue and public opinion? Answering these questions involves using a statistical procedure called causal mediation analysis,Footnote ¹⁴⁰ which has three steps. The first step estimates the effect of the treatment (i.e., NATO) on each mediator. The second step estimates the effect of each mediator on the outcome, support for war. The third and final step uses statistical information from the first two steps in a simulation process to estimate how much the treatment’s effect on the outcome is channeled through each mediator.

Figure 10 displays the results from steps one and two. The left panel shows the effect of NATO’s cue on each mediator. The estimated probit coefficients are reported in terms of predicted probabilities. When NATO recommends intervention, people are more likely to believe that intervention benefits U.S. status, will attract group participation, and will follow liberal humanitarian norms. Additionally, NATO’s policy cues make people more likely to believe that intervention can prevent crisis contagion, while it does not significantly change people’s beliefs about the chances of international retaliation and casualties. The right panel illustrates each mediator’s effect on people’s support for intervention. The mediators are not experimentally varied, so I include control variables to reduce confounding from omitted variables.Footnote ¹⁴¹ Each mediator affects people’s support for humanitarian intervention but to different degrees. Concerns about status have the most considerable effect, while fears of retaliation have the smallest effect.

Figure 10 Potential mediators that explain NATO’s effect. This figure gives the predicted marginal effects from probit coefficient estimates. The left panel plots the NATO effect on six mediators, and the right panel plots those mediators’ effect on people’s intervention support, controlling for potential confounds. The treatment, mediator, and outcome variables are binary. Estimates from the control variables are not displayed. Each of the 12 estimates is obtained from a separate regression (N=685). 95% confidence intervals are given. Data are from Survey USA-1.

Combining these two steps allows me to estimate each mediator’s average causal mediation effect (ACME),Footnote ¹⁴² or the extent to which each mediator can explain why NATO’s cues influence public opinion on war. Table 2 reports each ACME. Recall Section 3 in which Survey USA-1 revealed that NATO raises public support by about 26.4 pp. Beginning at the top of the table, the analysis shows that 1.5 of those 26.4 pp. are channeled through people’s changing beliefs about whether intervention would be consistent with humanitarian objectives and thus pro-norm.Footnote ¹⁴³ The results show that the other two social mechanisms – group participation and status/image considerations – are also substantial and statistically significant. NATO’s effect on public opinion also owes to its ability to shift beliefs about whether the humanitarian crisis would spread (Prevent Contagion), which is not predicted by the social cue theory, nor is it contradictory to the theory.

Meanwhile, people’s expectations about international punishment and casualties are not significant mediators.Footnote ¹⁴⁴ The last row reports the amount of NATO’s effect left unexplained by the mediation analysis: its overall 26.4 pp. effect minus the sum of all six mediation effects. Recall that the social cue theory argues that NATO exerts direct social pressure on ingroup members: presumably, some of this direct pressure is captured by the 9.2 pp. left unexplained by the causal mediation model, but this interpretation cannot be tested explicitly. In any case, the mediation analysis directly supports the social cue theory’s interpretation of legitimacy as alleviating relational concerns.

4.3 Test #3: Ruling Out Non-Social Mechanisms

This final test shows that NATO’s endorsement effect remains even after explicitly accounting for “objective” cost-benefit factors, such as the number of lives that would be saved, the financial and human costs of military action, and the mode of intervention. To rule out these alternative factors, Survey USA-4 included the original experiment in which NATO supports or opposes military action (NATO), while additionally randomizing whether respondents received or did not receive information about military action’s material cost and benefits. Specifically, those who received information read that “[m]ilitary action would save the lives of about [80 thousand OR 620 thousand] civilians. The operation would cost the U.S. government about [$850 million OR $4.1 billion], but the U.S. would avoid risking casualties by not sending ground troops.” Those who did not receive the information received no additional text.

If NATO’s cue raises public support for intervention by providing information about the material consequences of military action, directly providing that information to respondents should reduce the effect of NATO. However, the analysis is summarized in Figure 11 and shows that informing people about the material costs of intervention does not substantially reduce the effect of NATO’s cues. When the information treatment holds constant the financial and human cost of intervention, NATO shifts public opinion by 26.9 pp. When these factors are not fixed, the effect of NATO drops slightly but insubstantially to 25.7 pp.Footnote ¹⁴⁵ Thus, these material burden-sharing factors cannot fully explain NATO’s effect on public opinion.Footnote ¹⁴⁶

Figure 11 Material factors (financial costs and anticipated casualties) do not explain NATO’s effect. This figure reports the effect of NATO on public support for intervention depending on whether survey takers received information about the material costs of intervention (i.e., whether the material costs are “fixed”). N = 766. Data are from Survey USA-4.

5.1 Japanese Public Opinion on U.S. Intervention

This section shows how cues from the liberal community and NATO influence Japanese support for U.S. intervention. But before doing so, I elaborate on the political and theoretical relevance of the Japan case.

5.1.3 Effect of Social Cues on Japanese Public Opinion

The results of the two Japanese surveys replicate the main findings from the U.S. study. First, as Figure 12 reports, NATO’s policy endorsement affects Japanese support for U.S. humanitarian intervention more strongly than the Security Council. As the left-side figure shows, a sole Security Council cue increases public approval: the increase from Both Oppose to UNSC Only is about 7 percentage points (pp.). However, the more explicit ingroup cue by NATO raises support by a larger amount. The increase from Both Oppose to NATO Only is about 16 pp. Furthermore, obtaining the approval of both IOs does not significantly change public approval above and beyond just obtaining the approval of NATO (comparing NATO Only to Both Endorse). Turning to the right-side figure, the analysis also shows that the average NATO treatment effect is about 9 pp. higher than the average Security Council treatment effect. As with the U.S. surveys, the fact that the Security Council has a negligible effect conditional on receiving the ingroup cue from NATO, but not the reverse, cannot be easily explained by existing rational information theories.

Figure 12 NATO has a greater effect than the Security Council (Japan). The figure on the left shows public approval of U.S. intervention in four scenarios regarding NATO and the Security Council’s stance on intervention. The figure on the right reports the average treatment effect of each IO. 95% confidence intervals are displayed. N = 7,852. Data are from Survey JPN-1 & JPN-2.

Next, the results reported in Figure 13 show that the liberal community is more influential when sending its social cues through NATO. Specifically, the liberal community’s endorsement effect is about 3 pp. greater when channeled through NATO (comparing the top two estimates). While the effect size is modest, it is statistically significant and replicates the findings from the U.S. study. Next, due to the greater number of observations available in the Japan surveys, I could assess the effect of institutionalized cues from the broader international community, as represented by the Security Council. My theory argues that institutions can clarify the social meaning of a cue, but the international community does not represent any specific social group. Indeed, as Figure 13 shows, institutionalizing the international community’s endorsement vis-à-vis the Security Council does not change its effect on Japanese public opinion (comparing the bottom two estimates). Thus, institutionalized social cues are especially influential; institutionalized cues from no particular social group are not.

Figure 13 The effect of institutionalized cues (Japan). This figure shows the effect of the democratic community, the broader international community, and their cues sent through NATO or the Security Council on public opinion. N = 7,852. Data are from Survey JPN-1 & JPN-2.

5.2 Foreign Elites: A Survey of UK Parliamentarians

Social cues influence not just everyday citizens but also political elites. After all, elites are also social beings who care about norms, group belonging, and status. In fact, political elites might be expected to be even more conscientious about their country’s adherence to international norms and standing among peer nations compared to everyday citizens who are preoccupied with bread-and-butter issues. Indeed, existing research about social identity dynamics in international relations emphasizes state and elite-level dynamics.Footnote ¹⁶⁴ Earlier arguments about the signaling effects of IOs also highlighted the role of foreign elites.Footnote ¹⁶⁵

I examine political elites by surveying members of parliament (MPs) in the United Kingdom House of Commons.Footnote ¹⁶⁶ The United Kingdom is a critical player in the politics of humanitarian intervention, and its MPs are often directly involved in high-level policymaking, including holding relevant cabinet minister positions. Specifically, I contracted YouGov to poll a representative sample of 103 MPs.Footnote ¹⁶⁷ YouGov’s fieldwork took place in March and April 2023. While it was not possible to conduct a survey experiment on the MPs,Footnote ¹⁶⁸ I did survey a representative sample of them on a question relevant to the social cue theory. The question was the following:

In essence, this question asks respondents to compare policy options that mirror the UNSC Only and NATO Only prompts from the experimental studies. The social cue theory predicts that the MPs would favor NATO approval without the Security Council over the reverse because NATO approval represents a clearer ingroup cue. Figure 14 summarizes the results of this survey question and confirms the social cue theory’s prediction. Twenty-one percent of the MPs prefer humanitarian intervention with the Security Council but not NATO’s approval, while 47 percent prefer NATO but not Security Council approval. That is a 26 pp. difference.

Figure 14 When forced to choose, UK MPs prefer intervention with NATO over the Security Council. This figure shows the percentage of MPs in the UK Parliament who would rather support humanitarian with the Security Council but not NATO’s endorsement (UNSC Only) versus NATO but not the Security Council’s endorsement (NATO Only), or stated they “Don’t know.” N = 103. Data from Survey UK-MP.

While this statistic alone cannot be interpreted as causal or revealing why such preference exists, it does help to triangulate an additional dimension of the social cue theory in conjunction with other parts of this Element. We can also look to secondary historical evidence to corroborate NATO’s social impact on elites. For example, Gheciu (Reference Gheciu2005) shows how Romania’s entry into NATO socialized its policymakers to become amenable to liberal democratic rhetorical frames and policy arguments. This socialization process ultimately led Romanian lawmakers to accept arguments by the West regarding how Europe should address the 1999 Kosovo conflict. Thus, the survey shows that MPs in the United Kingdom generally prioritize NATO’s endorsement over the Security Council’s, and existing research documents social dynamics emanating from NATO to policymakers in other contexts.

5.3 Egypt, Outside the Liberal Community

To test some final implications of the social theory, I conducted a survey experiment outside the democratic context: Egypt. The Egypt survey mirrors the U.S. and Japanese surveys, except it simultaneously presented Egyptian survey takers with the policy position of three different IOs: the Security Council, NATO, and the Arab League. So, each IO was independently randomized to either approve or disapprove of humanitarian intervention, creating six different experimental conditions. Like the Japan survey, the dependent variable is support for U.S. humanitarian intervention. The survey was fielded by Qualtrics in Arabic to a diverse sample of 1,839 Egyptians.

The social cue theory implies that people will care first and foremost about ingroup cues. Thus, in the aggregate, the Arab League’s cue should have the most substantial effect on Egyptians because it represents an Arabic regional and cultural identity. NATO may also affect Egyptian mass opinion if some respondents identify with the liberal community, perhaps due to Egypt’s recent struggles for democracy, especially after the Arab Spring. But this prediction is less concrete ex ante. Lastly, the Security Council may or may not affect Egyptian public opinion. If it does, it would be for reasons outside the social cue logic.

The data are consistent with the argument that only ingroup cues influence mass policy preferences in Egypt. Figure 15 shows that neither NATO nor the Security Council shifts Egyptian approval to a statistically significant degree. However, the Arab League’s policy endorsement significantly increases Egyptian approval of U.S. humanitarian intervention by almost 6 pp. While this analysis does not assess causal mediators to explain the main experimental effects, it does show how the logic of ingroup cues operates in a different context than the liberal democratic community.

Figure 15 In Egypt, only the Arab League’s cue significantly affects public opinion. This figure shows the effect of each IO, averaging across the possible conditions of the other two IOs, on public approval of U.S. intervention. N = 1,839. Data are from Survey EGY.

6.1 Reframing Existing Evidence

Existing evidence has been interpreted as demonstrating the Security Council’s importance in granting legal legitimacy and in transmitting information about the material outcomes of war. How can we reconcile such interpretations with this study? In theory, both NATO and the Security Council could exert influence, albeit for different reasons. Additionally, as the Security Council includes key leaders from the democratic community, it may exert a social cue effect, though perhaps in a more diluted form. However, the evidence presented in this Element does not support these interpretations. Furthermore, upon closer scrutiny, previous studies can also be reinterpreted as being consistent with the social cue theory or at least inconclusive in differentiating between the competing explanations.

To begin, studies using historical data to examine the effect of international organizations (IOs) on public opinion have proven inconclusive, as there is limited historical variation in the institutional design and membership of these IOs.Footnote ¹⁶⁹ The Security Council has consistently held its role as the sole legal authority on war, governed by an elite pact, and structured as an independent, diverse, and conservative institution. This continuity provides little to no observational variation to test the mechanisms (i.e., elite pact, independence, etc.) proposed by existing theories. Additionally, cases in which NATO and the Security Council explicitly disagree are rare, further limiting the historical data’s ability to illuminate cross-institutional differences. From a social scientific perspective, these limitations make it challenging to use historical data alone to evaluate which institutions – and what institutional properties – truly matter.

Recognizing these limitations, some studies have turned to experimental data from public opinion surveys, as does this study. Two existing experimental studies could be interpreted as supporting conventional wisdom, but upon closer examination, one can be consistent with both conventional perspectives and the social cue theory, while the other provides direct support for the social cue theory. First, Grieco et al. (Reference Grieco, Gelpi, Reifler and Feaver2011) conducted a national survey experiment in the United States via telephone, finding that the joint endorsement of NATO and the Security Council raises American approval by approximately 24–27 percentage points (pp.). This effect mirrors the roughly 30 pp. effect I discovered in Survey USA-1 when comparing Both Oppose with Both Endorse, as discussed in Section 3. However, as Grieco et al (Reference Grieco, Gelpi, Reifler and Feaver2011) randomized both IOs simultaneously, it does not measure each IO’s independent effects.

Second, Tago and Ikeda (Reference Tago and Ikeda2015) tested Japanese public approval for U.S. military action under four conditions: (1) unanimous Security Council approval, (2) Security Council approval with abstentions by Russia and China and four nonpermanent members voting no, (3) lack of Security Council approval due to a veto by Russia and China, and (4) no attempt to seek Security Council approval due to anticipated opposition. While public support was higher in the first three conditions than in the fourth, support levels among those first three conditions were indistinguishable from one another. The authors interpreted this as an “A for effort” effect, suggesting that governments receive public approval mainly for attempting to secure Security Council authorization. However, social cue theory offers an alternative interpretation: Japanese approval increases so long as the United States secures the approval of democratic allies like the UK and France, while opposition from outgroup countries such as China and Russia is discounted. With these studies reconsidered through the lens of the social cue theory, I now directly reexamine additional claims made by the conventional theories.

6.2 Reexamining Legal Theories

One existing perspective argues that institutions legitimize armed interventions by legalizing them. To elaborate, since the creation of the United Nations, international law bans international wars with only three exceptions: self-defense, collective self-defense, and uses of force authorized by the Security Council.Footnote ¹⁷⁰ Under this system, the Security Council has broad authority to legalize the use of armed force to address, as a last resort, threats to international peace and security. Over time, the Security Council has expanded its mandate to authorize military interventions in humanitarian crises. Procedurally, the Security Council does this by passing a “Chapter VII” resolution, which requires nine of the fifteen members’ affirmative votes and no dissenting votes from the permanent five members, China, France, Russia, the United Kingdom, and the United States. Under this doctrine, ad hoc coalitions and alternative IOs like NATO do not have the same authority to legalize war.

International law could change how people think about war for several reasons. A Security Council resolution could convince people that a crisis poses a “threat to international peace and security” that can only be addressed via a military operation. But even if people are unfamiliar with the Security Council’s technical mandate, they could more generally prefer legal policies over illegal ones. This preference could be for normative reasons, as people worldwide favor legality, especially those residing in countries with a legalistic culture that respects the rule of law.Footnote ¹⁷¹ But such a preference could also be for more instrumental reasons. For example, Guzman (Reference Guzman2008) finds that abiding by the law helps to avoid reciprocity, reputation costs, and retaliation.Footnote ¹⁷² Thus, the Security Council could grant “legal legitimacy” to armed intervention,Footnote ¹⁷³ which could attract supporters for these various normative and instrumental reasons.

The previous sections showed how the Security Council, the only institution with the authority to legalize humanitarian war, exerts a modest influence on domestic and international opinion. This finding challenges the legal perspective, though proponents of this view may still raise two critiques regarding the analysis.

First, the aggregate mass public lacks an appreciation for the importance of international law, as most people are not well-informed about it.Footnote ¹⁷⁴ To address this critique, I demonstrate that the effect of IOs on public opinion is similar regardless of respondents’ understanding of international law. In Survey USA-1, respondents were asked: “Under international law, which of the following organizations do you think can authorize the use of military force in another country? Please select all that apply.” The response options were the following: United Nations Security Council, North Atlantic Treaty Organization (NATO), International Court of Justice (ICJ), Global Military Council, and None of the above.Footnote ¹⁷⁵ I categorized the respondents into “Know Law” and “Don’t Know Law” groups based on whether they correctly identified the Security Council as the sole authorizing body. Less than a quarter of respondents (237 out of 1,000) fell into the “Know Law” category.

Using this data, I examine whether respondents’ knowledge of international law influences how they respond to cues from the Security Council and NATO. Figure 16 reveals no such influence. Even those who correctly identified the Security Council as the sole legal authority remain susceptible to NATO’s social cues (as seen in the shift from Both Oppose to NATO Only along the solid line). While knowledge of international law slightly reduces public support for interventions lacking Security Council approval (indicated by the lower position of the solid line compared to the dotted line across Both Oppose and NATO Only), this reduction is modest at best.

Figure 16 Beliefs about international law do not moderate NATO’s Influence. Using probit coefficient estimates, this figure shows the predicted probability of supporting intervention for each treatment group, conditional on whether respondents know international law (N = 237) or do not (N = 763). Those who “Know Law” correctly identified the UNSC as the sole IO that can authorize interventions. 95% confidence intervals are displayed. These data are from Survey USA-1.

A second critique might argue that the true impact of international law on public opinion only emerges when respondents are directly informed about legal requirements. This perspective suggests an “out of sight, out of mind” effect for international law or implies that the law’s influence primarily manifests through public discourse, enabling political opponents to argue that a country has violated international law if it acts without Security Council approval. If this critique holds, NATO’s endorsement should have less impact on respondents’ approval when they are explicitly informed that Security Council approval is legally required. Specifically, informing people that a humanitarian intervention requires Security Council authorization should diminish NATO’s independent effect on public opinion.

I test this hypothesis using data from Survey USA-4. In this survey, I independently randomize (1) whether respondents receive the Both Oppose or the NATO Only scenario and (2) whether respondents receive an explicit statement about international law. All respondents read: “The United States is considering taking military action to help civilians in this crisis.” Respondents selected to receive the explicit statement about international law additionally read: “To be legal under international law, taking military action in this situation requires United Nations Security Council authorization.” Figure 17 visualizes the results, showing that NATO approval increases public support for intervention regardless of whether respondents were informed about international law (dotted or solid line).

Figure 17 Explicit information about international law does not suppress NATO’s influence. Using probit coefficient estimates, this figure shows the predicted probability of supporting intervention for each treatment group, conditional on whether respondents are (experimentally) informed about international law (N = 375) or not (N = 386). 95% confidence intervals are displayed. The data are from Survey USA-4.

However, providing legal information does reduce public approval in scenarios where both NATO and the Security Council oppose intervention. When respondents are informed of international law’s requirements, they are less supportive of the Both Oppose condition, suggesting that knowledge of legal standards increases sensitivity to Security Council disapproval – but only when NATO also disapproves. Once NATO endorses intervention, legal concerns largely disappear from consideration. This finding aligns with other research suggesting that international law influences public opinion primarily when it aligns with deeper norms and values. For example, Chu (Reference Chu2019) shows that international humanitarian law restrains public support for torturing enemy prisoners of war only if the opposing country also respects such laws; otherwise, support for torture remains unaffected by legal prohibitions. Similarly, Bayram (Reference Bayram2017) demonstrates that the impact of international law depends on respondents already holding cosmopolitan values. Thus, international law often reinforces preexisting norms, values, and identities – such as reciprocity and cosmopolitanism – rather than independently shaping public opinion.

6.3 Reexamining Rational Information Transmission Theories

I now turn to rational information transmission theories. Because there are several strands of this theory, I will first outline their key arguments before presenting additional analyses to evaluate them. These theories generally suggest that the Security Council, due to its unique institutional properties, conveys critical information about the costs, benefits, and likely consequences of military action through its policy endorsements.

One version of this argument views the Security Council as an “elite pact” whose approval can reassure foreign governments and citizens about the consequences of using military force.Footnote ¹⁷⁶ Specifically, Security Council authorization signals that military intervention will not lead to great power conflict or destabilize the international system, which can mitigate opposition to intervention. The Security Council’s distinctive attributes contribute to this perceived reassurance. First, its foundation as an institution brokered by post-WWII great powers means its decisions reflect the will of major players in global affairs. Second, the Council’s history of checking powerful states, including the United States, has established it as a credible institution for coordinating global security. Consequently, as some scholars have argued, the Security Council is “a more viable candidate than alternative institutions,”Footnote ¹⁷⁷ including NATO.Footnote ¹⁷⁸

A second perspective within the rational information transmission theories posits that independent or neutral IOs are best suited to relay useful information to observers.Footnote ¹⁷⁹ Drawing from formal models on the informational role of legislative committees,Footnote ¹⁸⁰ this perspective argues that IOs with a more heterogeneous or diverse set of veto players are more independent and, therefore, more influential. The logic here is that if such a diverse group of countries can agree on a policy, people will perceive such a policy as producing good consequences or at least not harmful. Focusing on how the Security Council, Thompson (Reference Thompson2009) notes that “in security matters, the Security Council … comes closest to operating as a neutral representative of the international community in a case of military intervention” and that “this logic helps explain why regional organizations, comprised of a less diverse set of states, do not produce a legitimation effect equivalent to that of the Security Council,” including “the more parochial NATO.”Footnote ¹⁸¹

A final related argument holds that the policy positions of conservative IOs provide the most informative signal to observers.Footnote ¹⁸² Conservative IOs, which rarely endorse military intervention or pose a “high legislative hurdle” to war, are seen as more credible when they do support it. In such cases, the “legislative hurdle” often stems from a pivotal voting member with dovish preferences or preferences that deviate most substantially from the intervening party.Footnote ¹⁸³ Therefore, when such a conservative IO endorses war, observers should be more likely to conclude that war would produce favorable outcomes. Again, this line of argument predicts that the Security Council’s blessing can raise support for war, especially in the case of humanitarian intervention, where veto-holders like China and Russia hold conservative preferences relative to the liberal community.Footnote ¹⁸⁴ In contrast, policy endorsements from more hawkish IOs that do not pose as high of a hurdle to authorization, like NATO, should be in an inferior position to relay helpful information. The logic is that if a warmongering institution supports war, it does not say much about whether war is a good idea.

The evidence thus far, however, already contrasts with these expectations. Sections 3 and 5 show that the Security Council had little impact once people in the United States and Japan received an ingroup endorsement from NATO. UK policymakers even preferred an intervention backed solely by NATO over one backed solely by the Security Council. Moreover, Section 4 employed a variety of research designs to show that the institutional cues are not operating through the logic of shaping people’s expectations about material outcomes. Thus, the Security Council’s endorsements are not consistently influential, nor can the domestic influence of IOs be entirely explained by people’s attempts to glean information about war’s material cost and benefits.

Based on this discussion, I evaluate three potential defenses for the rational information transmission theories. The first defense is another political ignorance critique, suggesting that much of the public does not hold “correct” perceptions about the Security Council and NATO. So, the problem is not with the theory but whether the theory generalizes to a largely ignorant public. Several pieces of evidence from previous sections contradict this argument. Specifically, the politically engaged public responds to the cue of IOs in ways that resemble the mass public: college-educated Americans who are registered to vote and frequently follow international political news respond to ingroup cues perhaps even more acutely than the mass public (discussed in Section 3 and reported in the Online Appendix). This result also aligns with the finding reported in Section 5 about the UK MPs, who are knowledgeable policy elites that should be likely to “correctly” perceive the Security Council’s institutional properties.

A second defense might be that there are just too many differences between NATO and the Security Council to isolate the effect of these specific institutional properties on public attitudes. Section 4 presented several pieces of evidence to show that NATO’s effect can be attributed to social cueing; however, these findings are not necessarily direct negative evidence for the alternative mechanisms relating to institutional independence and conservativeness.

Here, I directly analyze the rational informational transmission theories’ central assumptions about independence and conservativeness, which existing approaches operationalize using the diversity and distance of preferences of an institution’s member countries.Footnote ¹⁸⁵ Specifically, in the second Japan survey, I measured each respondent’s beliefs about a selection of NATO and the Security Council’s key member countries in terms of whether respondents believed that key members from each IO are committed to human rights (norms) and, alternatively, politically aligned with Japan (strategic).Footnote ¹⁸⁶ The individual responses were then used to calculate an independence score (i.e., the heterogeneity/diversity of ratings, as measured by the variance of ratings), and a conservative score (i.e., the most distant rating or “highest legislative hurdle,” using the strategic measure only). For space, the Online Appendix discusses these research procedures in greater detail.

The results are clear. As summarized in Table 5, the general public perceives the Security Council as far more independent and conservative than NATO. Thus, the Security Council’s overall weak cue effect cannot be explained by mass ignorance about its institutional properties. To delve deeper, using multivariate regression analysis for individual-level analysis, I estimated how perceiving each IO as independent or conservative affects the ability of that IO to change people’s support for humanitarian intervention. As summarized in the final row of Table 6, the interaction coefficient IO*Property (which can be either independence or conservativeness) is generally positive, as the conventional wisdom would predict. That is, viewing the Security Council or NATO as more independent or conservative amplifies the effect of its cues on public opinion. However, the size of that effect is substantively small, and it is also statistically indistinguishable from having zero effect.

The third potential defense is that IOs are intended to influence only a specific subset of people – namely, those who are skeptical of an intervening country’s political motives and concerned about the potential consequences of war for the international system.Footnote ¹⁸⁷ According to this view, the Security Council’s legitimizing effect should be most potent among individuals who perceive the interventionist country as reckless or imprudent in international affairs. If IOs serve this reassurance function, then those who doubt the intervening country’s reliability and restraint most strongly should be most responsive to Security Council endorsements.

I test this argument regarding the effect of IOs among war-weary citizens, again using data from the Japanese public opinion on U.S. intervention. The survey takers answered the question, “Do you agree or disagree with the following statement: ‘The U.S. generally makes good decisions about using military force in other countries.’” Depending on their answers, respondents were grouped into three categories: those who agreed, those who neither agreed nor disagreed, and those who disagreed. These three groups were labeled as those who believe the United States to be Prudent, Neutral, or Imprudent.

Figure 18 displays the effect of IOs among these groups. Contrary to the reassurance thesis, the more skeptical the audience is of U.S. military power, the more they turn to NATO for political reassurance. Regardless of whether the respondent viewed America as a prudent, neutral, or imprudent military power, the effect of a Security Council cue is about 2–3 pp. In contrast, the impact of NATO grows with the survey respondent’s increasing belief that the United States is an irresponsible military power, peaking at 14 pp. among the most skeptical respondents. This trend implies that NATO can play a reassurance role. However, the effect of a NATO cue is about 9 pp. even among those least skeptical of the United States (i.e., those who believe the United States is prudent), implying that NATO’s cue is not only about reassuring skeptical audiences.

Figure 18 NATO, not the Security Council, reassures skeptics of the United States. This figure shows the percentage point effect of the Security Council and NATO on Japanese approval of U.S. intervention, conditional on whether the survey respondents view the United States as a prudent, neutral, or imprudent military power. N = 6,090. 95% confidence intervals are given. Data are from Survey JPN1 & 2.

Material and power-centric perspectives cannot make sense of these findings. If people only held material concerns about unchecked military power in the international system, it would make little sense to seek reassurance from NATO, an organization often associated with a warmongering West. Instead, the social cue theory can demystify these results. Japanese citizens who are skeptical of the United States will look to the liberal community and NATO for reassurance, and even if they do not seek reassurance, they still will respond to social pressures to support the United States when faced with an ingroup cue. To summarize, this section showed how past evidence can be reframed as consistent with the social cue theory, and it further presented new evidence that raises questions about the core assumptions made by existing theories of IOs.

7.1 Where the Scholarship Now Stands

This Element theorizes a new perspective on how legitimization functions in political behavior, particularly by developing, applying, and testing the social cue theory to understand how international institutions legitimize foreign policy. While international relations scholars have recognized the political relevance of legitimacy since at least the 1960s,Footnote ¹⁸⁹ the realist view that dominated through the Cold War largely dismissed institutions as mere reflections of power dynamics, either paralyzed like the Security Council or simply reflecting pre-existing geopolitical competition like NATO. However, as the Cold War waned and global cooperation through international institutions expanded, scholars began developing new accounts of why international institutions matter. Some of them, particularly those writing from a constructivist approach, returned to the elusive idea of legitimacy and better specified what it means and how it influences politics at the state and system levels.Footnote ¹⁹⁰ Subsequently, rationalist scholars reframed legitimization as a process in which institutions, due to their institutional design and membership composition, transmit information about how domestic audiences weigh the costs, benefits, and motives behind a country’s foreign policy actions.Footnote ¹⁹¹ This Element builds upon these insights about legitimacy and how institutions channel their influence through domestic politics while innovating in theory and evidence.

Specifically, this Element elucidates a new causal model of legitimization from the perspectives of social identity and political psychology. The social cue theory explains how political communities and institutions influence group members by exerting ingroup pressure and alleviating social-relational concerns about norms abidance, group behavior, and status and image. Applied to the phenomenon of humanitarian intervention, the argument implies that the liberal community and NATO vis-à-vis their liberal identity influences the citizens and elites of democracies, the primary participants of these human rights operations. The preceding sections substantiated these claims while ruling out alternative explanations relating to burden sharing, NATO’s military strength, mass ignorance, and the liberal community’s conflation with Western regionalism.

In advancing a new argument on why international institutions matter, this Element also highlights the limitations of existing perspectives rooted in international law and rational information transmission. For example, the Security Council’s authority to legalize actions does not substantially shift public opinion in the presence of ingroup cues, even when audiences are informed of relevant international laws.Footnote ¹⁹² Additionally, the ability of institutions to exert influence does not appear to depend on public perceptions of their independence or conservativeness, nor can their surprising effects be explained by widespread political ignorance. Nonetheless, the social cue theory does build on the established insight that institutions can shape domestic political views, influencing the views of citizens and elites worldwide.

7.2 The Liberal Community and Humanitarian War

Beyond advancing theoretical debates, understanding how international institutions legitimize war has significant practical ramifications, especially for governments weighing the role of multilateral institutions in their foreign policy. The trade-offs are complex. Institutions can confer legitimacy, mobilizing domestic and international policy support, but they can also restrict a country’s autonomy and decision-making freedom. Some institutions like the Security Council represent a broader range of the international community, but they are prone to gridlock, which could forestall urgent action to alleviate mass human suffering. Because of these trade-offs, it is essential for policymakers to understand which types of institutions yield what types of effects, whether they be social or material, and why.

This Element does not directly address the ethical debates on humanitarian intervention but offers insights relevant to them. Unilateral actions remain widely unpopular, as seen when President Obama scaled back plans for Syrian intervention in 2013 due to a lack of institutional backing. Yet, as this study suggests, governments seeking legitimacy for their actions may not always need Security Council approval. Instead, political endorsements from ingroup peers and institutions, especially within the liberal democratic community, often confers much of the same legitimizing power. For those advocating greater restraint in international politics, this finding may be troubling: governments can sidestep broad multilateral institutions like the UN in favor of institutions within like-minded communities, reshaping the scope of international interventions.

This Element also highlights a broader question for liberal democracies. As authoritarianism rises and democratic norms wane, institutions like NATO face new pressures and critiques from within. Leaders who question the value of these institutions, often for their perceived costs, may overlook their broader role in sustaining public support for foreign policy and maintaining global stability. While the liberal democratic community has stood together for over half a century, its survival is not guaranteed. Norms, identities, and institutions can erode over time, and this erosion would come at a great cost – not only in lost identity and purpose but also in the institutional tools that have proven central in managing international challenges.