International organizations (IOs) have increased in prominence over the past several decades. From universal bodies such as the United Nations (UN) and the World Trade Organization (WTO), to regional bodies like the European Union (EU), IOs shape policies and outcomes across a wide range of issue areas. Yet the effectiveness and legitimacy of these organizations ultimately depend on the support of domestic political actors in member states. This special issue of Review of International Organizations presents new perspectives on the complex interplay between domestic politics and IOs.

The significance of this topic is underscored by the mounting challenges faced by IOs in recent years. In 2016, the United States (US) began obstructing appointments to the WTO’s dispute settlement mechanism. In 2018, under a different administration, the US officially withdrew from UNESCO and then exited from the Paris Climate Agreement in 2020. During the COVID-19 pandemic, the US announced its intention to withdraw from the World Health Organization. While some of these policies were subsequently reversed, the US continues to paralyze the WTO, maintains high tariffs on Chinese goods, and pursues industrial policies that provide large subsidies to domestic industries. As of the writing of this article, neither political party in the United States supports trade agreements with market access provisions, and a majority of Americans (59%) believe that the US has lost more than it has gained from increased trade.^Footnote1

The backlash to international organizations is not limited to the United States. Europe, Brazil, and India, among others, have pursued reforms in international organizations to reassert their autonomy. Under President Jair Bolsonaro, Brazil challenged the European Union on environmental issues and sought to reshape its role in the United Nations. In India, Prime Minister Narendra Modi has emphasized the importance of respecting national sovereignty and territorial integrity, particularly in digital governance, a stance that contrasts with the approach of international organizations, which advocate for more borderless regulation. Threats to IOs have also emerged from non-populist sources. Japan withdrew from the International Whaling Commission in 2019, a decision made by a mainstream political party that reflected Japan’s long-standing pro-whaling stance. Similarly, the United Kingdom’s withdrawal from the European Union (Brexit) in 2020 was executed by a mainstream conservative party. Between 1945 and 2014, there were more than 200 cases where member states voluntarily withdrew themselves from all contractual obligations and legally terminated their memberships in IOs (von Borzyskowski & Vabulas, 2019, 2024b). This crisis of international organizations brings to light a long-standing fundamental tension between the push for greater global cooperation and the desire for nations to prioritize domestic interests and national sovereignty (e.g., Rodrik, 2012; De Vries, 2018; De Vries et al., 2021; Heinkelmann-Wild et al., 2024; Lake et al., 2021; Bowen & Broz, 2022). To explain support and opposition for international organizations, we need a deeper understanding of the individual domestic political actors and the domestic institutions that constrain the choices of policy-makers in dealing with IOs.

In this introduction to the Special Issue, we present a stylized theoretical model to help organize the existing literature on international organizations, including contributions to this Special Issue. At its core, our model examines the dynamics of IOs through the lens of individuals within nations—individuals that often hold differing, and sometimes conflicting, preferences about the policies that IOs address. Differences arise not only from diverse priorities but also from unequal access to resources. Some individuals may rely solely on their votes, while others can wield greater influence through organized interest groups, campaign donations, or lobbying. Political institutions, which evolve over time, play a crucial role in channeling these preferences into policy outcomes. These institutions—ranging from laws governing lobbying and elections to voter access—are shaped by the deeper, more stable rules embedded in domestic constitutions. Constitutions, such as a country’s system of government, are persistent, vary across nations, and influence how citizen preferences are expressed and represented in IO decision-making.

We distill the interaction between citizens, governments, and other countries into the Citizen’s IO game. We show that the international IO policy choices of domestic coalitions are equivalent to an outcome in which each citizen selects welfare weights to assign to each citizen across the globe. Welfare weights are essentially a way to account for the varying interests of citizens across nations. An allocation of welfare weights to non-domestic citizens implies IO membership. Welfare weights are balanced across individuals in different countries to maximize personal welfare while minimizing negative international externalities. For an IO to exist and adopt a policy, there must be a system of welfare weights that supports that policy in equilibrium. To illustrate the main features of our model, we provide a simple example focused on trade policy. Using our framework, we show that the standard non-cooperative outcome of trade-policy settings implies an outcome with no IO members. We also show the impossibility of fully internalizing global externalities through IOs when sovereignty is respected and citizens are treated equally.

We are certainly not the first to theorize about the domestic politics of international cooperation and international organizations (for foundational works, see for example Viner, 1950; Gourevitch, 1978; Milner & Rosendorff, 1997; Moravcsik, 1998; Putnam, 1988; Frieden, 2002). However, we present a micro-level approach to complement the existing literature, which often employs a more ‘macro’ perspective. For instance, research on the geopolitics of international organizations frequently abstracts from individuals to nation-states, treating these as unitary actors that pursue a broadly defined ‘national interest’ (Martin, 1992; Thacker, 1999; Oatley & Yackee, 2004; Barro & Lee, 2005; Vreeland, 2008; Lake, 2009; Dreher et al., 2009; Schneider, 2009; Copelovitch, 2010; Davis, 2023; Carnegie & Clark, 2023). In contrast, our micro-level approach emphasizes that the foreign policy actions of sovereign governments—including their decisions to participate in and shape international organizations—are deeply rooted in the preferences and priorities of individuals, which are then aggregated through both domestic and international political institutions. By focusing on this individual-level foundation, our approach provides a more nuanced perspective that supplements existing theories. In particular, our approach enables a more detailed examination of the mechanisms by which individual preferences—shaped by factors such as economic interests, political ideology, and social identity—translate into national foreign policy and, ultimately, into international outcomes. Our model demonstrates that citizens are ultimately at the center of many of the questions that scholars of international organizations have debated for decades. This approach not only helps organize the existing literature on international organizations but also opens up new avenues for research on the micro-foundations of global governance, the role of sovereignty in international cooperation, and the design of more effective international institutions.

We present a stylized theoretical model to assist in organizing the extant literature connecting IOs and domestic politics. The model highlights the role of citizens within each country in shaping their country’s support for, and participation in, IOs. The formation of policy in domestic and international institutions can be seen as the direct outcome of citizen preferences for policy. This echoes classic work in political economy using citizen-candidate models of domestic policy-making (e.g., Osborne & Slivinski, 1996). Other scholars have explored the link between citizen preferences and IOs, often focusing on levels of aggregation of citizen actors—such as electorally relevant citizens, political parties, or interest groups—or different mechanisms through which citizen preferences aggregate to the IO level.^Footnote2 The model we present shows that the main forces connecting preferences to international policy simplifies to maximizing a weighted global welfare function in a manner similar to Lindbeck and Weibull (1987). In our model, the choice of welfare weights is endogenous, but has a status quo bias. This reflects important features of domestic and international politics, and illuminates various links in the chain connecting citizens to IO policy.

Domestic constitutions play a crucial role in constraining how citizen preferences are aggregated into both domestic and international policies. By setting the boundaries of governmental power and delineating the roles of the legislative, executive, and judicial branches, constitutions create procedural constraints that influence whether and how citizen preferences shape foreign policy. For instance, some constitutions require legislative approval or public referenda for international agreements, thereby limiting executive autonomy and ensuring broader political alignment. At the same time, constitutional principles—such as federalism, sovereignty, or individual rights—limit the scope of permissible domestic policies, reinforcing citizens’ preferences around autonomy or social values. Constitutions thus serve as frameworks that mediate the influence of citizens on policy, channeling preferences through structured pathways and imposing constraints on government action. This also includes systems of government.^Footnote3 We provide a simple way to formulate constitutional constraints as the allowable set of welfare weights in the weighted global welfare maximization problem of each citizen. As domestic politics is an endogenous object, the formulation also allows us to capture the feedback loop from policy formed within IOs to the changes in domestic and international politics.^Footnote4

The model begins intentionally general, and IOs serve the role of coordinating on policy to reduce a negative international externality. This allows us to capture much of the existing work. After laying out the general framework, we turn to a special case of multi-dimensional policies that take values that have real numbers—such as a tariff or defense spending—and well-behaved preferences, which reflects many settings of interest. This allows a general characterization of the citizens’ trade-offs, and highlights the balance between retaining domestic control over policy and working with other countries to reduce international externalities (Ikenberry, 2000; Stone, 2011; Ikenberry, 2011; Rodrik, 2012; von Borzyskowski & Vabulas, 2019; Heinkelmann-Wild et al., 2024; Lake et al., 2021; Kreuder-Sonnen & Rittberger, 2022). In Section 7, we apply the framework to the trade policy setting game in Bowen et al. (2025).

The international community

The international community consists of a group of J countries . Each Country j is simply a set of citizens  who interact repeatedly over an infinite horizon . The citizens of each country are guided by their domestic constitution to control domestic policy. Denote Country j’s domestic policy in period t by .

Domestic policy  can include, for example, subsidies (Bagwell & Staiger, 2006; Rickard, 2018), domestic transfers (Cameron, 1978; Katzenstein, 1985; Rickard, 2012), monetary policy (Clarida et al., 2002; Eichengreen, 2013; Bechtel et al., 2014; Broz, 2015; Bordo, 2021), tariffs (Grossman & Helpman, 1995; Bagwell & Staiger, 1999; Maggi & Rodríguez-Clare, 2007), financial contributions to IOs (Broz & Hawes, 2006; Broz, 2008; Schneider & Urpelainen, 2014; Schneider & Tobin, 2016; Reinsberg, 2023), democratization (Pevehouse, 2002, 2005; Poast & Urpelainen, 2015; Blair et al., 2023), autocratic stabilization (Debre, 2022; Cottiero, 2023; Cottiero et al., 2024), immigration policy (Betts, 2011; Peters, 2017; Johns et al., 2022; Alesina & Tabellini, 2024), environmental policy (Bernauer, 1995; McLean & Stone, 2012; Genovese, 2019; Tingley & Tomz, 2022), fiscal capacity (Besley, 2020), and other foreign policies (Voeten, 2005; Chapman, 2009). Note that we do not place any restriction on the policy space  at this time, so,  can be multidimensional and include non-numeric variables. The vector of policies in all countries in period t is denoted .

Resources

Citizen i in Country j has resources  in period t from which they derive instantaneous utility , where . Resources can include wealth, income, human capital, other forms of capital, etc. Writing the utility function in this way does not preclude a more micro-founded model with consumption and prices, as these micro-foundations typically rely on citizens making trade-offs to deploy resources.^Footnote5

We assume resources of Citizen i in Country j are given by a function , where  is a country-level parameter that can change over time, shifting the resource distribution in Country j. This parameter could include, for example, national debt, demographics, capital stock, labor supply, technology, or natural resources (Alesina & Drazen, 1991). It can be, therefore, shaped by policy, subject to random shocks or related to past values of the parameter. Let , so the resource distribution parameter  depends on the previous period’s distribution through , the international vector of policies , and a random shock . The random shock can include, for example, natural disasters, technological innovations, or budget shocks.^Footnote6

Note that the resource distribution in Country j can be affected by their own policies , but also by policies in other countries , i.e., . For example, a tariff imposed by the United States affects producers of that good in countries across the world. Or a decrease in development assistance from the European Union can have implications for the poor in Guinea-Bissau. Countries can, therefore, impose externalities on each other through their domestic policies, which creates a role for IOs (Nordhaus, 1975; Sandler & Tschirrhart, 1980; Nordhaus, 1994, 2015; Ostrom, 2015). To make this dependence more explicit, we focus on each agent’s expected indirect utility from their own domestic policy and the domestic policy of other countries, and denote this by

With indirect utilities written in this way, it is clear that policy-setting is, in fact, a game between citizens across all countries. Many papers studying international policy-making consider it as a game between countries behaving as unitary actors, as for example depicted in typical bargaining models of international cooperation (e.g., Fearon, 1998). As stated earlier, one purpose of this model is to peel back the layer of “government,” and instead show how IOs and international policy are direct results of citizen preferences and choices. We describe domestic policy-making next, and then introduce its connection to IOs.

Domestic policy-making

A substantial literature connects domestic politics, and elections in particular, to IOs (Dreher, 2003; Dreher & Vaubel, 2004; Caraway et al., 2012; Schneider, 2013; Rickard & Caraway, 2014; Chaudoin, 2014; Schneider & Slantchev, 2018; Schneider & Tobin, 2020; Schneider, 2019b, 2020). We borrow language from the classical collective choice literature (e.g., Arrow, 1950, 1963; Austen-Smith & Banks, 1996) to formulate politics and policy-making as the outcome of actions by a domestic decisive coalition. We say that Country j’s policies in period t are selected by a period t decisive coalition . A decisive coalition  is a set of citizens who can influence policy. It can include the identity of specific policy makers, such as senators or congressmen, veto players, such as a president or executive, and non-veto players, such as a majority of citizens. The decisive coalition  might also include members based on resources , such as firms that make campaign contributions, or lobby groups that support various policies. The key feature of a decisive coalition in Country j is that it is a collection of citizens from Country j, each with their own indirect utility from policy .

We assume the set of possible decisive coalitions in period t is determined by Country j’s period  policy choice , i.e., . The choice of  can include, for example, changes in campaign contribution limits that give more or less power to the wealthy, laws related to lobbying, or voter registration policies (Milner & Rosendorff, 1997; Broz & Hawes, 2006; McLean, 2015; Genovese, 2019; Brutger, 2024). Such policies allow different coalitions without identifying specific members. It can also include elections of public officials, if allowed by the constitution. Note that, since  can specify if a citizen is in the coalition based on resources , policies of all countries can affect the decisive coalition in Country j through their influence on Country j’s resource distribution parameter . For example, OPEC members deciding how much to produce may affect elections or changes of governing coalitions in Country j, through changes in the international price of oil (Arezki et al., 2022). Economic policy decisions in other countries can also directly affect electoral outcomes (Kayser, 2007, 2009; Kayser & Peress, 2012; Walter et al., 2018) or economic conditions (Forbes & Warnock, 2012; Rey, 2013; Bruno & Shin, 2015; Miranda-Agrippino & Rey, 2020) in other countries.

The allowable set of decisive coalitions is constrained by a country’s constitution. Country j’s constitution is assumed to be fixed through time, and specifies the system of government, frequency of elections (if any), party formation, parliamentary versus majoritarian, if laws are governed by simple majority, super majority, or other. For example, a constitutional democracy cannot disenfranchise all citizens, whereas a monarchy or authoritarian regime can allow a single citizen to be the sole member of the decisive coalition. The constitution may not specify individual citizens who are enfranchised, but may describe features of an enfranchised citizen (such as gender or race). The specific individuals in the decisive coalition are constrained by those features. Thus, while the constitution is fixed through time, the members of the decisive coalition need not be. In practice, even in a dictatorship, there can be other informal influences on the policy choice, such as advisors, public opinion, or international pressure (Gandhi & Przeworski, 2007; Weeks, 2008; Magaloni, 2008; Svolik, 2012; Fang & Stone, 2012; Boix & Svolik, 2013; Weiss, 2013; Hashimoto, 2020). The scope of this influence can be captured by different features of a constitution. Formally, we denote Country j’s constitution by  and assume . We postpone the discussion of how decisive coalitions are formed to Section 6. We first describe how IOs are modeled and their relationship to decisive coalitions in each country.

International Organizations

We assume an international organization serves the role of coordinating policies across countries to reduce transaction costs and manage international externalities (Nordhaus, 1975; Keohane, 1984; Nordhaus, 1994; Kaul et al., 1999; Barrett, 2003; Schneider & Slantchev, 2013; Ostrom, 2015; Nordhaus, 2015).^Footnote7 As discussed above, international externalities are created when the policies of one or more countries have spillover effects on citizens in other countries (Sandler, 2004). These spillovers can be unintended and are often negative. International externalities can arise in various contexts, such as civil and interstate conflict, climate change, environmental degradation, global health issues, financial instability, trade protection, migration crises, terrorism, organized crime, and cybersecurity. Externalities cross national borders and require collective international efforts to manage.

We also model an IO as a collection of citizens who coordinate to make decisions about their own domestic policies that produce international externalities. We denote the set of individuals in the IO in period t by . Individuals in the IO are themselves members of a country in the international community, so we denote by  the set of IO members from Country j in period t. In line with common definitions of IOs as formal entities that possess a permanent secretariat and have sovereign states as members (Singer & Wallace, 1970; Abbott & Snidal, 1998; Pevehouse et al., 2019), we assume that IO members from Country j must be members of the decisive coalition in Country j, and thus . Similar to domestic coalitions, we assume that the previous period’s domestic policy in Country j determines the Country j members of the IO in period t, so we write this as . The collection of individuals who participate in the IO in period t are the individual members from each country in period t, so . Thus, individuals in an IO may change as domestic coalitions and their designated IO representatives change.

Timing

The timing of the policy-making game between citizens is as follows. At the beginning of period t, members of the domestic coalition and IOs are in place for each Country j (i.e.,  and ). The previous period’s resources are also known and given by . Domestic policies  are selected by  and  for all j through a process to be described in the next section. The random shock  is realized, and together with  and , determines the current period’s resources for each agent in each country . Citizens then derive utility  from resources. Finally, depending on policy choice  the next period’s decisive coalitions  and IO members  are selected.^Footnote8

Observation 1

Policy  affects the following variables:

1. 1.Period t resource distribution parameter in each country: ;
2. 2.Period  set of allowable decisive coalitions in each country, ;
3. 3.Period  IO members from each country, .

We can now describe decision-making within domestic decisive coalitions and IOs.

For each Country j, the resource distribution, possible decisive coalitions, and IO members (, , and ) are inherited from the previous period, and thus describe the state of the country at the beginning of period t. Formally, the state of Country j in period t is , and the state of the world is . We assume citizens discount the future by a factor , and  denotes Citizen i’s expected discounted infinite horizon payoff when policy vector  is chosen in state . To ease notation, we will henceforth denote period t variables without t subscript, and period  variables with a prime:

In each period, given the state of the world s, each citizen in Country j seeks to maximize their expected discounted payoff .

We assume that coalition member i in Country j can only directly affect their domestic policy  whether or not they are a member of the IO. This assumption is in line with the fact that most IO policies have to be ratified and implemented by sovereign governments at the national level. For example, decisions in the WTO to reduce tariffs have to be translated into domestic policy, using the political decision-making processes as prescribed in national constitutions. Similarly, if NATO members collectively agree to increase collective security measures, individual member states have to get approval for greater defense spending from the appropriate national political bodies. Even decisions of international courts have to be translated into policy by national governments.^Footnote9 In the European Union, for example, Franchino (2004) finds that over 87% of legislative acts delegate implementation to national governments.^Footnote10

Citizens maximize their individual expected discounted payoff through the choice of domestic policy , according to their role in the domestic decisive coalition  and possibly their role in the international organization . These policy choices are made conditional on the policy choices of other countries . All citizens in a decisive coalition must account for the incentives of other citizens in the coalition when selecting policy . Members of IOs are distinguished by the fact that they account for, not only their domestic coalition incentives, but also other IO member’s payoffs when choosing domestic policy . This allows for IO members to coordinate their domestic policy as part of the equilibrium of the game, including a system of punishments for off equilibrium path behavior. This is the standard view of policy cooperation in IOs via repeated games (Snidal, 1985; Oye, 1985; Fearon, 1998; Martin & Simmons, 1998; Koremenos et al., 2001; Rosendorff & Milner, 2001; Rosendorff, 2005). The difference in our approach is the players in the game are the individual citizens of each country, versus the country as a whole. Moreover, since domestic coalitions include non-governmental members through, for example, special interest groups, it implies that non-governmental actors can be included in IOs. The US-Mexico-Canada Agreement (USMCA) Secretariat, for example, consists of members of each country, including governmental and non-governmental actors.

We incorporate these considerations by assuming that for Citizen i to agree to any set of policies, Citizen i’s expected dynamic payoff  must exceed some individual threshold . This threshold can be determined by, for example, a status quo policy, the citizen’s bargaining strength, wealth, citizen’s “power,” IO decisive rules, or other features. The threshold for each citizen  plays an important role in the analysis, and we refer to them as demands. Various strands of the literature across economics, political science, and international relations have discussed factors that affect these demands, and we summarize these in Section 4. For coalition members not in an IO, the thresholds of other countries do not enter their maximization problem, but for IO members, the thresholds of other IO members do. We close this section with the theoretical formulation of the policy choice problem that accounts for these demands.

Consider the maximization problem of Citizen i in Country j, who is a member of the domestic coalition , and also a member of . Assume (for now) that Citizen i is an agenda setter within the IO and the domestic coalition. Citizen i seeks to maximize her own payoff conditional on satisfying the other members of her domestic coalition and the members of the IO. Citizen i’s problem is thus

The first set of constraints Eq. C-D includes one for each member of i’s domestic coalition, and the second set in Eq. C-IO includes one constraint for each member of the IO from each of the other countries. This problem captures the international and domestic pressures faced by any agenda-setter in an IO—they arise from the demands of each member of the IO and each member of the domestic coalition—highlighting the tension between satisfying domestic constituents while maintaining international commitments.

The maximization problem in Eq. 1 also captures the tension between maintaining sovereignty and solving international collective action problems. Consider a Citizen m in Country k who is a member of k’s domestic coalition but not a member of the IO. This citizen faces only her analogous constraints in Eq. C-D. While this may seem advantageous for m because there are fewer constraints, Citizen m loses the ability to coordinate with other countries on their domestic policies to improve aggregate payoffs in a way that most benefits m. Citizen m would not have a “seat at the table.” It is entirely possible to free-ride on the coordination of other citizens in Country k, but in some cases, coordination by other citizens can have a negative effect on m. This is exemplified by the well-documented (Autor et al., 2013) “China shock” which caused higher unemployment, lower labor force participation, and reduced wages for individuals in local labor markets that housed import-competing manufacturing industries. Similarly, all citizens in Country k may free-ride on the policy coordination by citizens of other countries in the IO, but those other citizen choices can have a negative effect on citizens from Country k. Aggregating this to the country level gives us the familiar argument of countries outside an IO free-riding on the coordination of international policies by countries in the IO (Gruber, 2000; Schneider & Slantchev, 2013).

We highlight that participation in the IO requires balancing the needs of domestic and international constituents, and hence giving up some sovereignty. These trade-offs are discussed in a large literature that looks at the use of majority decision-making rules in international organizations and the delegation of power to international IO bureaucrats (Pollack, 1997; Barnett & Finnemore, 1999; Tallberg, 2002; Hug, 2003; Nielson & Tierney, 2003; Hawkins et al., 2006; Vaubel, 2006; Voeten, 2007; Lake, 2007; Voeten, 2008; Haftel, 2013; Johnson, 2013a, b; Johnson & Urpelainen, 2014; Johnson, 2017; Honig, 2019; Clark & Zucker, 2024). Papers in this area have shown that international IO bureaucrats may have preferences that diverge from their member state and have the power to shift policies toward their own goals. Scholars have debated the extent to which bureaucrats have individual influence on policy outcomes, but there is a growing consensus that IO staff can influence IO policies in many areas. Our framework captures this influence, and the divergent interests of individual members of the IOs.

Before discussing the solution to the agenda setter’s maximization problem in Eq. 1, we discuss the literature that has addressed the sources of demands  and  by coalition and IO members.

The demands of domestic coalition members and IO participants play a crucial role in shaping a country’s engagement with IOs. In the model, Citizen i in Country j faces two sets of constraints when setting domestic policy to maximize their payoff: (i) domestic Eq. C-D; and (ii) international Eq. C-IO. These constraints are determined by the demands of domestic coalition members , and IO members from other countries . As mentioned above, a citizen’s demand can be determined by numerous factors including status quo policy, a citizen’s bargaining strength, wealth, citizen’s “power,” IO decision rules, systems of government, laws, etc. We now discuss the literature that has contributed to our understanding of the domestic and international sources of these demands.

4.1 Domestic factors

There are many domestic factors that influence demands, but we focus on three themes: political institutions, resources, and public opinion. Political institutions are the endogenous rules of domestic political systems that aggregate citizen preferences, and essentially endow coalition members with de jure power in the sense of Acemoglu and Robinson (2008). They structure policy-making and channel the preferences of citizens regarding IO participation. This includes factors such as electoral systems (as specified in the constitution), veto players, the distribution of power between branches of government, and laws that allocate or delegate authority. Resources, as described in Section 2, can include wealth, income, human capital, other forms of capital, personal characteristics, the importance of the economic sector to which the citizen belongs or participation in interest groups within a country that exert pressure on policy-makers to pursue specific policies within IOs. Public opinion reflects the preferences and beliefs of the general public regarding international cooperation and specific IO policies. Public opinion can be shaped by factors such as culture, perceptions of international threats, economic conditions, politicization of IOs, elite messaging, or media portrayal of IOs. These features of domestic political institutions interact to determine the demands of domestic coalition members.

A substantial literature on domestic political economy emphasizes that the possibility of leader turnover in domestic political institutions significantly impacts the demands placed on decision-makers (Alesina & Tabellini, 1990; Azzimonti, 2011; Acemoglu et al., 2011). In democracies, electoral pressures and the need for public support constrain policy-makers’ actions. For example, during election periods, governments are more likely to push policies within IOs that align with domestic preferences, even if these may not be optimal from a global perspective (Dreher, 2003; Dreher & Vaubel, 2004; Caraway et al., 2012; Schneider, 2013; Conconi et al., 2014; Rickard & Caraway, 2014; Chaudoin, 2014; Schneider & Slantchev, 2018; Schneider & Tobin, 2020; Schneider, 2019b, 2020). In the model, the demands of voters would be greater in an election year than in a non-election year. In a democracy, a winning majority of voters is necessarily part of the domestic coalition. Similarly, an autocrat threatened by civil unrest must account for the demands of the citizens leading the unrest as they conduct international negotiations (Gandhi & Przeworski, 2007; Weeks, 2008; Magaloni, 2008; Svolik, 2012; Fang & Stone, 2012; Boix & Svolik, 2013; Weiss, 2013; Hashimoto, 2020).

Previous work shows that veto players in the domestic political system play a significant role in IOs (Schelling, 1960; Putnam, 1988; Milner & Rosendorff, 1997; Tarar, 2001; König & Finke, 2007; Mansfield et al., 2000, 2002, 2008; Milner & Mansfield, 2012; McLean & Stone, 2012). The demand of a citizen who is a veto player must be met for any domestic policy to be changed, and ultimately, for any international policy to be changed. A veto player could be a median voter in a democracy or the President. Countries with more veto players have greater domestic coalitions and must incorporate the demands of each veto player in international negotiations. This can lead to policy outcomes that are biased towards IO members facing more domestic veto players’ constraints (Ali et al., 2023). The presence of veto players can thereby impact a country’s ability to engage in international cooperation and the depth of integration pursued in those agreements (Downs & Rocke, 1995; Fearon, 1998; Mansfield et al., 2008; Mansfield & Milner, 2012).

The nature of the electoral system can influence which citizens are prioritized. For example, Rogowski and Kayser (2002) build from a Stigler-Peltzman framework and find a bias that favors consumers in majoritarian systems and favors producers in proportional representation systems. This consumer bias in majoritarian systems could have implications for how countries engage with IOs. Countries with majoritarian systems might be more inclined to push for trade liberalization in WTO negotiations, as it aligns with consumer interests for lower prices. However, other scholars predict the opposite. Grossman and Helpman (2005) develop a model that envisages a protectionist bias in majoritarian systems.^Footnote11 Rickard (2022) also shows that countries with PR systems, party-centered electoral competition, and successful green parties tend to lead efforts to restrict environmentally damaging fishing subsidies, i.e., by increasing their demands for these policies. The demands imposed by domestic coalition members may depend on the extent to which adopted policies are binding domestically, and the extent to which they are monitored and enforced (Snidal, 1985; Oye, 1986; Martin & Simmons, 1998; Maggi & Rodriguez-Clare, 1998; Abbott & Snidal, 2000; Donno, 2010; Hafner-Burton et al., 2015, 2023, 2024).^Footnote12

Domestic resource distribution also significantly shapes the preferences of and demands placed on policy-makers regarding IO participation (Osgood & Kim, 2019; Lee & Osgood, 2019; Osgood, 2021; Méndez & Van Patten, 2025). Firms and industry groups often lobby governments to pursue specific policies within IOs that benefit their sectors (Pelc, 2009; Dür et al., 2015; McLean, 2015; Bechtel et al., 2019; Genovese, 2019; Brutger, 2024). The extent of this influence often hinges on the resources each group of citizens can use to lobby or persuade (Grossman & Helpman, 1994, 1995). For example, different interest groups have shaped trade policy at various points in history. Policy-makers must often balance the interests of import-competing sectors, which may favor protectionist policies, against those of export-oriented industries that benefit from trade liberalization (Grossman & Helpman, 1994; Beaulieu, 2002; Hicks et al., 2014; Milner & Tingley, 2015; Alt & Gilligan, 1994; Bailey et al., 1997; Irwin, 2017; Bowen et al., 2025). Existing research also shows that citizens organized through interest groups have a strong impact on cooperation within IOs and compliance with IO mandates (Milner, 1988; Dai, 2005; Simmons, 2009; Chaudoin & Urpelainen, 2015). The effect of citizens goes beyond substantive IO policies, but can also affect, for example, decisions of the IO to accept new member states (Schneider, 2007, 2009; Schneider & Urpelainen, 2012; Davis & Wilf, 2017; Davis, 2023).

Since coordination in the IO affects individuals’ payoffs directly, it shapes and is shaped by public attitudes toward IOs (Bechtel & Tosun, 2009; Bechtel et al., 2017; Walter et al., 2018; Zvobgo, 2019; Dellmuth et al., 2019; Dellmuth & Tallberg, 2021; De Vries et al., 2021; Madsen et al., 2022; Jurado et al., 2022; Tingley & Tomz, 2022; Brutger & Clark, 2023). As citizens form attitudes toward IOs based on their perceived performance and impact, these views reverberate through domestic political processes and affect the demands of domestic coalition and IO members (Risse-Kappen, 1991). Voters consider IO performance when making electoral decisions, and governments respond by adjusting their policies toward IOs to align with voter preferences, especially during election periods (Dreher, 2003; Dreher & Vaubel, 2004; Schneider, 2013, 2019a, b, 2020; Rickard & Caraway, 2014; Chaudoin, 2014; Caraway et al., 2012).^Footnote13

Of course, public opinion is influenced by various factors beyond the actions of the IO, including nationalism, economic conditions, media portrayal of IOs, and the popularity of a government (Gabel & Palmer, 1995; Gabel, 1998; Hobolt & Vries, 2016). Nationalist sentiments can lead to skepticism or opposition toward IOs perceived as infringing on national sovereignty (Copelovitch & Pevehouse, 2019). Economic conditions, such as unemployment or income inequality, may shape public attitudes on trade agreements or financial assistance programs administered by IOs (Mansfield & Mutz, 2009; Bechtel et al., 2014; Walter, 2021).^Footnote14 Media coverage also plays a role in framing IO activities and shaping public perceptions (Baum & Potter, 2019; Jurado et al., 2022; Brutger & Clark, 2023). During periods of nationalist fervor, media narratives can activate national stereotypes and influence how the public processes information about IOs and foreign affairs (De Vries et al., 2021). The impact of public opinion is particularly evident in democracies, where electoral pressure and the need for public support constrain policy-makers’ actions regarding IO participation. For example, situations in which a United States president enjoys high approval ratings and security at home, or one whose party controls both the House and the Senate, contribute to a higher demand from their IO representatives (Lohmann & O’Halloran, 1994; Milner & Rosendorff, 1997; Karol, 2000).

Two studies in this Special Issue relate to the impact of public opinion on coalition and IO member demands. Hunter and Walter (2025) introduce a new dataset containing over 600,000 statements about international organizations (IOs) from six national parliaments between 1990 and 2018 to show that members of national parliaments speak more negatively about international organizations when IOs are perceived to have too much authority or a democratic deficit. In another contribution to this Special Issue, Schlipphak et al. (2024) use a survey experiment conducted in six countries to show that these effects are mitigated by individual identity. In particular, citizens’ cosmopolitan identity interacts with the authority of international organizations (IOs) to shape citizens’ domestic support for international organizations. Citizens with higher cosmopolitan identity are more likely to support IOs that hold greater authority, while those with lower cosmopolitan identity become more skeptical of IOs with increasing authority.

The interplay of domestic political institutions, resources, and public opinion on IOs is explored in a contribution to this Special Issue by Lee and Lee (2024). Lee and Lee (2024) demonstrate the significant role of domestic concerns, such as voter preferences and partisan priorities, in influencing governments’ responsiveness to international disputes. The study explores how the partisan orientation of governments in democratic countries influences their decisions to settle investor-state disputes (ISDS) arising from regulatory changes. Using both quantitative analysis of 335 ISDS cases (1994 - 2020) and qualitative case studies, it demonstrates that right-leaning governments are more likely to settle disputes informally, prioritizing economic growth and foreign investment, while left-leaning governments prefer arbitration to protect domestic social welfare and regulatory autonomy. These preferences are further shaped by domestic political constraints, with legislative power amplifying the partisan differences in settlement behavior.

Interactions such as these determine the domestic constraints and opportunities faced by policy-makers when engaging with IOs, and influence the extent and nature of a country’s participation in global governance.

4.2 International factors

Citizens who are members of international organizations face a complex array of factors imposed by the IO that influence their demands. These include the formal design of international organizations—variations in veto rules, consensus, majority voting—the informal limitations on individual members’ ability to shape policy, membership in other IOs or international coalitions, personal career concerns, and even personal financial or political benefit. Scholars have increasingly recognized the importance of the preferences and priorities of IO member states, the design of international organizations and how they influence decision-making outcomes at the international and national level (Gray, 2009, 2013; Hafner-Burton & Schneider, 2019; Ferry et al., 2020; Debre, 2022; Cottiero & Haggard, 2023; Cottiero et al., 2024; Koremenos et al., 2001; Dür & Mateo, 2010; Hooghe & Marks, 2015). We discuss how these affect IO member demands.

A significant factor influencing IO member demands comes from the formal rules governing IOs. These rules vary widely, from unanimity or consensus-based decision-making to majority voting systems (Blake & Payton, 2015; Hooghe & Marks, 2015; Hooghe et al., 2016). Some IOs—such as the UN General Assembly or the World Trade Organization—give every member country one vote; other IOs—such as the UN Security Council, the International Monetary Fund, or the Council of the European Union—use weighted voting schemes where some member states receive more votes than other countries, based on some predefined formula (oftentimes related to the relative size of the member’s contributions, its GDP, or population). These formal rules allow individuals in IOs more or less voting weight—formal rules that give an individual more voting weight increases their demand and less voting weight decreases it. Members are less constrained under unanimity rules, as they can block decisions that are not in their favor. Conversely, majority voting systems can impose constraints on individual members, as decisions may be made despite their opposition (Schneider, 2011; Steinberg, 2002; Schneider & Urpelainen, 2012; Lewis, 2003; Haftel & Thompson, 2006; Achen, 2006; Thomson et al., 2006).^Footnote15

Holding leadership positions within international organizations—such as chairing intergovernmental bodies or rotating into decision-making bodies—can provide members with agenda-setting power, further influencing outcomes (Baron & Ferejohn, 1989; Tallberg, 2006; Ali et al., 2023); and some states have increased their influence as they rotate into important decision-making bodies within the institution (Dreher et al., 2014, 2016; Vreeland & Dreher, 2014). Finally, the sovereignty costs of member states increase when IOs have greater enforcement capacity (Hafner-Burton et al., 2015, 2023). This will raise the demands of IO members.

Beyond formal voting power, informal power dynamics also play a crucial role.^Footnote16 The correlation between structural economic or military power and vote shares in many IOs has led to a veritable cottage industry of scholarly work demonstrating that structurally more powerful states have a significantly greater ability to shape IO policies according to their interests (Thacker, 1999; Abbott & Snidal, 2000; Stone, 2002, 2008, 2011; Vreeland, 2003; Broz & Hawes, 2006; Tallberg, 2008; Arregui & Thomson, 2009; Dreher et al., 2009; Aksoy, 2010; Copelovitch, 2010; Morse & Coggins, 2024). More powerful states often wield influence through informal channels, inducing other states to align their votes accordingly (Dreher & Jensen, 2007; Dreher et al., 2009; Vreeland & Dreher, 2014; Carter & Stone, 2015). For example, the US enjoys significant voting power in the IMF, correlating with its economic clout, leading to policies that often reflect US preferences beyond what one would expect on the basis of the formal voting rules. Studies have shown how such power asymmetries can influence IO effectiveness (Stone, 2008; Lipscy & Lee, 2019). Informal power can also be exercised more indirectly, for example when the IO staff is composed mainly of nationals of a particular member state (Voeten, 2008; Thomson & Dumont, 2022; Lang et al., 2024).

Of course, voting power and decision-making rules can interact with each other, and other important factors further influence international constraints. For example, decision-making becomes more difficult as member states’ preferences diverge (Kahler, 1992; Alesina & Spolaore, 1997; Alesina et al., 2001; Gilligan, 2004; Schneider & Urpelainen, 2014; Schneider & Tobin, 2016), and this effect is exacerbated when decisions are made by unanimity rule. The constraints induced by preference heterogeneity has become especially salient in recent years when illiberal regimes in democracies and autocracies have increasingly acceded and participated in international organizations (Brazys & Dukalskis, 2017). To the extent that these regimes have different priorities and preferences than their liberal counterparts, it has had significant effects on member constraints overall (Cottiero et al., 2024; Meyerrose & Noorudd 2025; Winzen, 2023). While vote shares of individual members matter, the ability of less powerful states to form voting coalitions (Lyne et al., 2009; Schneider & Tobin, 2013; Rickard, 2022), use veto power opportunities (Schneider, 2011), or link issues (McKibben & Western, 2014) can affect individual members’ demands significantly.

Last, the increasing complexity of the IO landscape constitutes another constraint. In many policy areas, there exist numerous IOs dealing with similar or overlapping goals. Whereas the jury is still out on whether the existence of alternative IOs can create positive synergy effects or lead to competitive pressures, there is increasing evidence that the existence of competing IOs can shift pressures and priorities, thereby affecting IO member demands (Busch, 2007; Lipscy, 2015; Pratt, 2018; Kinne & Kang, 2023; Qian et al., 2023).

We will show that the solution to Citizen i’s maximization problem in Eq. 1 is equivalent to the solution of the dual problem of maximizing some weighted global welfare function. The approach used is similar to Lindbeck and Weibull (1987) and others who applied this to domestic policy-making. We demonstrate the importance of this formulation for coalition and IO member choice, and highlight the trade-offs involved with inclusion of various coalition and IO members.

To keep the mathematics straightforward while maintaining the intuition, we assume policies may be multidimensional, but must take numeric values. Specifically, we now assume . A typical example is a single-dimensional policy such as a tariff or subsidy. We further assume that each citizen’s dynamic payoff  is differentiable and strictly concave in  for all . We present the discussion with a single-dimensional policy to keep notation light, but the extension to multiple dimensions is immediate.

Let  be a weighted welfare function for Country j with welfare weight for coalition member n in Country k given by . Then

We can similarly construct weighted welfare functions  with associated welfare weights , for each country . Each country’s welfare function puts weight (potentially zero) on coalition members from each of the other countries in the IO. The equivalence theorem below connects this weighted welfare function with the maximization problem in Eq. 1.

Theorem

(IO and Weighted Global Welfare Equivalence). For each country , if a decisive coalition  chooses domestic policy , then there exists a set of welfare weights  such that

The proof is simple yet illustrative, so we discuss it here. First note that the maximizer of  is unique.^Footnote17 Now let i be the agenda setter for Country j and a member of . Then i chooses Country j domestic policy  to solve the maximization problem in Eq. 1. By standard arguments, Citizen i’s optimal solution satisfies the first order necessary condition of the Lagrangian

For some agent  from some country , the solution to this problem is the IO policy choice . The Lagrange multipliers for coalition member  in Country j, and IO member m from Country  are given by  and  respectively. Note that members of  are already included in . There is, thus, a Lagrange multiplier for each member of j’s domestic coalition and a Lagrange multiplier for each member of the IO from each of the other countries. The following remark highlights the first insight from our micro-founded approach.

Remark 1

The Lagrange multiplier  can be thought of as Citizen i’s shadow price of Citizen n being in the decisive coalition , and the Lagrange multiplier  is Citizen i’s shadow price of Citizen m from Country k being in the IO.

Remark 1 provides us with a way of conceptualizing (and ultimately quantifying) the cost to agenda setter i of the domestic coalition and IO members’ demands  and . If the coalition or IO member has a low demand, then that member’s shadow cost to i will be low, and if the demand is high, then the coalition member is more expensive to i. As discussed in Section 4, demands can arise from the IO structure, domestic pressures, individual power, among other factors. For example, if the IO member’s country is very powerful, it can be very costly to satisfy them, as they can sway policy far from what maximizes i’s payoff. If a coalition member has a particularly onerous domestic political constraint, such as a powerful lobby group or an upcoming election, then the demand is high, and the shadow price is high. All the factors discussed in Section 4 contribute to these shadow prices being greater or lower. Section 8 suggests future research to capture these demands and shadow prices.

Denote the full set of Citizen i’s Lagrange multipliers by . Each country’s agenda setter, for example Citizen m, will solve a problem similar to Eq. 1. They will each have a corresponding Lagrangian  and Lagrange multipliers . Country k’s agenda setter will thus account for demands of all IO members from each country (including j) when setting policy . In this way, all IO members from each country internalize the externality of other IO members when setting domestic policy, consequently reducing the global externality. This leads to our second insight from our micro-level approach.

Remark 2

The benefit of the IO is that each country in the IO internalizes the impact of their policy choices on the payoffs of other IO members through the constraints analogous to Eq. C-IO. This consequently reduces the global externality.

While many papers have considered IOs as reducing negative international externalities, we show that these externalities can be directly incorporated into each citizen’s payoffs through the impact of foreign country policies, i.e., through . When agenda setter i from Country j accounts for IO member m’s payoff in setting domestic policy , the result is reducing the negative externality caused by  on IO member m’s payoff from Country k. When IO member m additionally accounts for i’s payoff in the same manner, the consequence is improving the aggregate outcome.

This disaggregated approach shows how improvements in welfare can be uneven across citizens. This depends on the factors that influence the demands of coalition and IO members as discussed in Section 4. Moreover, the policies set within the IO are heavily dependent on the individual preferences of citizens who are members of the IO. These demands and preferences translate into the shadow prices of each IO member, leading to different welfare outcomes.

By the Kuhn-Tucker Theorem, Citizen i’s optimal policy  and the Lagrange multipliers for Citizen i’s problem (which are possibly zero) satisfy the following first order condition

A given set of coalition members  and IO members from each country give rise to a different vector of Lagrange multipliers , which are a direct result of the demands of each of these citizens.

We now return to the weighted welfare function . The maximizer  of  satisfies the first order necessary and sufficient condition

We construct welfare weights  as follows

With welfare weights given by , the policy  that satisfies the first order condition for the weighted welfare function  in Eq. 5, is the same as the solution to the first order condition to the Lagrangian in Eq. 4. This is because  is simply a renormalization of . The policy choice of a decisive coalition  must satisfy the agenda setter’s optimization problem for some , and thus the policy choice of coalition  is the same policy choice as a social planner maximizing the weighted welfare function , with welfare weights given by . This completes the proof.

The Theorem tells us that a set of welfare weights , can be rationalized by some decisive coalition  and some IO members with appropriate demands  and . This leads to our next major insight from this micro-founded approach.

Remark 3

A vector of welfare weights for each country  determines the optimal policy vector . This policy satisfies demands of some domestic coalition and some IO members. Moreover, a set of welfare weights for each country determines payoffs for citizens across each country, and these payoffs are unique.

Given the equivalence between the IO member problem and the weighted welfare problem, we recast the policy choice of coalition members and IO members (and their associated demands), as a choice of assigning welfare weights. By choosing a given set of welfare weights, it implies a particular set of optimal policies and payoffs. They can be rationalized by a particular set of domestic and IO coalition members, and their respective influence on policy through formal and informal sources. We can thus rewrite Citizen i’s expected dynamic payoff in period t as a function of , i.e.,

where . We therefore state Citizen (i, j)’s policy choice as a selection of welfare weights  that maximizes . The policy choice game between citizens across all countries is then a choice of welfare weights for each citizen in each country.

The state s is now the parameter  from the previous period that determines expected resources , and the previous period’s (endogenous) welfare weights , which capture the existing domestic coalitions and IO. The period’s resource parameter  factors into Citizen i’s decision about the welfare weights for the current period (and hence the coalition, IO members, and policy). If no new weights are agreed to, the default is to revert to the previous period’s weights. This is consistent with the fact that domestic laws are enforceable by courts and remain intact unless changed deliberately (or expire based on how they are written). This built-in status quo bias is related to a rich literature studying legislative bargaining with an endogenous status quo (Bowen et al., 2014; Anesi & Seidmann, 2015; Dziuda & Loeper, 2016; Ali et al., 2023). This is how current domestic coalitions and IOs affect future domestic coalitions and IO choices (and policies). Previous welfare weights capture the current “demands.”

As discussed in Section 4, demands are shaped by domestic and international forces. In the Citizens’ IO game these are now captured in part by exogenous factors—the constitution  and shocks to resource —and in part by endogenous factors, through the welfare weights that are now part of the state variable s. They determine future IO polices and payoffs. Other existing work provides evidence that aligns with these dynamics. In a contribution to this Special Issue, Tobin (2025) demonstrates how International Investment Agreements (IIAs) can influence individual actors’ perceptions and behaviors, illustrating the feedback between domestic politics, international organizations, and individual citizens (in this case, foreign investors) in member countries. Levelu et al. (2025) demonstrate this in a setting related to immigration. They find that visa provisions in regional trade agreements increase the bilateral stock of immigrants between the two countries. Moreover, the effectiveness of these provisions is reduced by anti-immigration sentiment of voters in the destination. This shows the impact of IO policy on public opinion.

The endogenous component of the state provides a framework to understand backlash against IOs (Lake et al., 2021; Gray, 2024). Backlash may manifest as government resistance or withdrawal due to diverging preferences between the domestic demands and the IO policy (von Borzyskowski & Vabulas, 2019; Broz et al., 2021). In those situations, member governments have incentives to try and affect the decision-making process to avoid unpopular outcomes (Winzen, 2023; Meyerrose & Nooruddin, 2025; Lipps & Jacob, 2024).^Footnote18 To the extent that this is not possible—such as in areas where they have given up significant sovereignty—the inability to satisfy the member constraints can potentially lead to the weakening or death of IOs (Gray, 2018; von Borzyskowski & Vabulas, 2024b).

Note that a domestic coalition member is now any citizen with a non-zero welfare weight under the status quo. Formally

We can now define the remaining elements of the citizens’ IO game more precisely.

Constitutions

The welfare weights of Citizen i in Country j must respect Country j’s constitution . In the citizens’ IO game we formalize the constitution as a restriction on the allowable set of welfare weights. For example, a democracy would not allow  that places full weight on Citizen i. Welfare weights with this feature would be excluded from . A monarchy or autocracy might allow for such vectors. Let  denote the infinite dimensional simplex that captures all possible combinations of welfare weights. Then it must be the case that welfare weights chosen by Citizen i in Country j satisfies .

Equilibrium

Let  be a strategy for Citizen i in Country j that assigns a welfare weight to each citizen of each country, given the state of the world s. In equilibrium these must be a best response to the welfare weights chosen by Citizen m in Country k. Let  be the profile of strategies for each citizen in each country. We write Citizen i’s expected dynamic payoff when the state is s and given strategy profile  as . We can now define an equilibrium of the citizens’ IO game.

Definition 1

An equilibrium of the citizen’s IO game is a strategy profile  such that

for all , ,  and s.

The condition in Definition 1 is the standard best response condition. The equilibrium  thus determines the equilibrium policy in Country j and IO membership. We assumed that if citizens of Country j cannot agree to a new domestic coalition, then the status quo coalition prevails. This ensures the existence of an equilibrium.

Proposition 1

The equilibrium policy  solves the following system of equations for all  and 

Proposition 1 characterizes the equilibrium policy as the result of the equilibrium welfare weights selected by each country. This will be a system of simultaneous equations, one for each country. Together, they determine all policies across countries. Necessarily, the choice of domestic policies must be consistent across all citizens in that country. While domestic welfare weights must be consistent, across countries welfare weights can vary. This is because the welfare weight of Country j’s citizens only matter for Country j’s domestic policy.

Proposition 2

In equilibrium, Citizen i from Country j is a member of an IO if  for some .

In equilibrium, any citizen placing positive welfare weight on citizens from other countries in policy selection, is effectively a member of the IO. For example, an authoritarian regime needs to give up some domestic policy making authority when it joins IOs. It does this by placing positive welfare on citizens from other countries. Note, all citizens in Country j must agree for this to be an equilibrium.

Consider two countries . Country 1 has an economy as in Bowen et al. (2025) (henceforth BBR) with two citizens  and two non-numeraire goods . Citizen 1 in Country 1 has current period resources  with the first element denoting endowment of good 1, and the second indicates good 2 endowment. Citizen 2 has resources . Assume  so that Citizen 1 is the exporter in Country 1. Country 2 has a single citizen , who has resources  which are his endowments of goods 1 and 2 respectively. The economy in Country 2 is therefore as in Maggi and Rodríguez-Clare (2007). To keep matters simple, in this example we assume  is fixed for all j, so that  for all j.

Domestic policies

Let policy  be Country j’s tariff (or subsidy if negative) on good . So the global policy is the two-dimensional vector . The global optimal tariff for a utilitarian planner would be zero, but due to terms-of-trade effects, citizens in each country have positive ideal tariffs. As in BBR, tariffs result in indirect utility . These payoffs are single peaked in  for citizens of Country j, and strictly decreasing in . They satisfy the assumptions in Section 5. Denote the ideal tariff for Citizens 1 and 2 in Country 1 as  and . We know from BBR that  and these are independent of  but depend on  and . Citizen 1 in Country 2 has ideal tariff  which is the standard optimal tariff in Maggi and Rodríguez-Clare (2007). We know from BBR that .

Constitutions

A set of welfare weights for all citizens in all countries is . The constitution of each country  determines the set of allowable welfare weights. One could image that the notion of “sovereignty” prescribes that each country’s citizens will carry a weight no less than citizens from other countries, and the total weight should also not be less than the total weight assigned to other countries. Let us denote this special class of constitutions as  and define it formally as

We assume  for all j. Country 2 is a dictatorship, and Citizen 1 can choose any allocation allowed that respects sovereignty, and thus . We assume Country 1 is a democracy, and thus some weight must be allocated to each citizen. The details of the constitution will determine if some citizens are systematically disadvantaged. For example, in Malaysia, the constitution provides special rights and privileges to ethnic Malays and other indigenous peoples but not to other citizens. In Israel, the Basic Laws (considered constitutional laws) distinguish between Jewish and non-Jewish citizens, particularly concerning rights to immigration under the Law of Return. For simplicity, we assume no such distinction, so the constitution of Country 2 treats citizens equally. However, equal treatment in the constitution does not guarantee equal treatment in practice. That is, equilibrium coalitions can have uneven weights. To capture this, if Country j is a democracy assume this implies an equal minimum weight to each of its citizens .^Footnote19 Thus we define the class of democratic constitutions  as

We assume Country 1 is a democracy, i.e., . A strategy  for Citizen i from Country j is a choice of welfare weights that satisfies the democracy and sovereignty constraints. Each citizen’s strategy specifies welfare weights for each citizen in each country, i.e., .

Uncooperative Equilibrium

Without cooperation, as a dictator, Citizen (1, 2) places full weight on his own welfare and thus . Citizen (1, 2) is the only domestic coalition member and is not a member of the IO because there are no citizens from other countries to whom he allocates welfare weights. By Proposition 1, the equilibrium domestic policy in Country 2 would be Citizen (1, 2)’s ideal, . Without cooperation, each citizen  in Country 1 would assign some weight greater than  to each other and none to Country 2 in equilibrium. That is , with  to respect the democracy constraint. This would result in an equilibrium policy between the two citizens’ ideal  as in BBR.^Footnote20 We summarize this below as a Corollary to Propositions 1 and 2.

Corollary 1

The uncooperative equilibrium in the two country game is

with . This results in equilibrium policies

and there are no IO members.

Cooperative Equilibrium

As is standard, we assume that by repeated game logic, citizens can sustain outcomes that improve on the non-cooperative outcome. IOs serve to coordinate this equilibrium. They facilitate communication, coordination, information gathering and dissemination, and monitoring and sanctioning, as is common in discussions of the theory of IOs (Fearon, 1998; Koremenos, 2001; Abbott & Snidal, 1998; Martin & Simmons, 1998; Oye, 1985; Snidal, 1985; Rosendorff, 2005; Fang, 2008). For example, in a contribution to this Special Issue, Mangini (2024) studies the US Generalized System of Preferences (GSP), which offers tariff-free access to US markets for developing countries conditional on their compliance with labor and intellectual property rights. The study shows that sanctioning threats are effective in motivating states to comply with the GSP’s conditions, even when they have not been directly confronted, as states adjust their policies preemptively to avoid losing market access. IO rulings, even when non-binding, can exert at least a short-term effect on domestic policy. This is also demonstrated by Schoner (2024) in another contribution to this Special Issue. The study focuses on “violation decisions” issued by UN treaty bodies, particularly the Human Rights Committee, which oversees the International Covenant on Civil and Political Rights. These decisions, while not binding, are found to exert short-term improvements in physical integrity rights (e.g., protection from torture and extrajudicial killings), especially when civil society organizations (CSOs) are involved in publicizing the outcomes.

Stationary grim trigger strategies can sustain cooperation, but a vast IO literature also shows that strategies can account for a changing state of the world as captured by  and .^Footnote21Among other things, this work shows that on the equilibrium path, an IO can allow periodic “escape” from commitments. Our framework helps to flesh out how these nuances are influenced by citizens, constitutions, and random shocks to resources. This can result in the IO collapsing on the equilibrium path, or never quite getting off the ground. The next result highlights such an instance.

In this simple economy the world welfare maximizing tariff is zero and would result from welfare weights (1/4, 1/4, 1/2). The question is if these weights are compatible with domestic institutions and domestic bargaining power of citizens. Sovereignty implies they are not because it violates the condition .

Proposition 3

Sovereignty implies that free trade with welfare weights (1/4, 1/4, 1/2), is not an equilibrium outcome in the Citizens’ IO Game.

Proposition 3 extends readily beyond this simple example. If countries insist on sovereignty (i.e., prioritizing national interests and citizens) and treat all citizens equally, then pure free trade is not possible.^Footnote22 More generally, global externalities cannot be fully internalized if countries abide by sovereignty and treat all citizens equally. The degree of cooperation will be limited by the constitution and the status quo. This echoes Rodrik’s Trilemma of global integration (Rodrik, 2012) and work by scholars who attribute instances of democratic erosion to the delegation of policy autonomy to international organizations (Meyerrose, 2020, 2024). The result also extends to other topics including the environment, financial risk mitigation, security arrangements, labor, etc. In light of Proposition 3, it should come as no surprise that countries routinely miss climate goals set by the UNFCCC (as an example).

We have presented an impossibility result, but it is more useful to determine what can happen in a cooperative equilibrium. This goes beyond the scope of this introduction, but is a great topic for future study.

We presented a theoretical framework that illuminates the complex interplay between domestic politics and international organizations. The discussion highlights the trade-offs involved in pursuing optimal policy outcomes through IOs. There is a balance between the individual benefits of internalizing a global externality, satisfying the demands of domestic and international coalitions embedded in the status quo, and respecting domestic constitutions.

The paper by Bondi et al. (2025) in this Special Issue exemplifies the connections between individual citizen preferences and IO performance that the model highlights, particularly regarding how domestic actors shape international cooperation through IOs. Specifically, Bondi et al. (2025) demonstrate how firms involved in global value chains (GVCs) push for deeper preferential trade agreements (PTAs), which aligns with the Citizens’ IO Game where individual preferences influence policy outcomes in international organizations. The findings illustrate how the welfare weights assigned to different actors in the theoretical model can be influenced by economic interests, such as those of firms engaged in GVC activities, leading to specific policy choices in international trade agreements. Our model also sheds light on the legitimacy gap between elites and citizens (Dellmuth & Tallberg, 2023; Dellmuth et al., 2022a; Dellmuth & Tallberg, 2015; Dellmuth et al., 2022b) as elites are more likely to belong to the coalition that is more able to shape IO policy outcomes and should therefore be more likely to benefit from IO policies.

Our framework provides several insights for future research. One fruitful avenue to explore empirically and theoretically, is how the domestic and international forces identified in Section 4 influence citizens’ demands in a more consistent framework. This would allow the influences to be more easily compared, and develop new insights on IO outcomes and sustainability. For instance, future studies could examine how IO institutional design features might reduce member constraints while accommodating varying domestic and individual considerations. Researchers might also investigate how adjustments in IO decision-making timelines could mitigate the impact of domestic political pressures, such as election cycles, on international negotiations and policy outcomes. Analyzing how different domestic political systems shape countries’ engagement with IOs and their policy preferences within these organizations would deepen our understanding of IO dynamics. One immediate theoretical question arising from Proposition 3 is whether welfare weights that treat citizens unequally can sustain optimal welfare outcomes while respecting sovereignty. More generally, what are the classes of welfare weights that can support optimal outcomes accounting for the endogenous nature of these weights? The answers could shed light on the policies that can be feasibly implemented through IOs. As also mentioned, there is contradicting evidence on the impact of majoritarian systems and the incentives for trade liberalization in WTO negotiations. One could use our framework for easier empirical comparison.

Another important theme to explore is the effect of sometimes opposing forces generated by exogenous economic conditions on IO member demands. Conventional wisdom suggests that poor economic conditions can generate backlash against IOs, whereas some scholars find that economic crises can increase incentives of governments to cooperate within IOs to address economic crises more effectively (Haftel & Nadel, 2024). Our model can guide scholars in comparing these trade-offs through examining how they influence IO member demands in different contexts.

The tension between sovereignty and global cooperation, as identified in our findings, also merits further exploration, especially in the context of increasing government emphasis on these issues. Future research could focus on innovative IO designs that better balance national sovereignty with the need to address global challenges. Studies could analyze how varying conceptualizations of sovereignty across countries affect their willingness to engage in deep cooperation through IOs. Investigating the long-term implications of the difficulty in fully internalizing global externalities while respecting national sovereignty would also be a valuable area of inquiry. Furthermore, analyzing the welfare implications of IO policies at the individual citizen level, rather than solely at the country level, could yield valuable insights.

The feedback loop of IOs to citizen preferences and political institutions is captured by the endogeneity of the status quo welfare weights in our model. Existing work can further explore this feedback at a disaggregated level. Work by Choi et al. (2024) is a groundbreaking example of such an endeavor in the context of trade policy. They robustly demonstrate that US counties whose 1990 employment depended on industries vulnerable to NAFTA suffered large and persistent employment losses after its implementation. Moreover, voters in these counties turned away from the party of President Clinton, who promoted the agreement. Further studies can apply similar techniques in other country and policy contexts, to make the links between status quo coalition members, IO formation and policy (as captured by status welfare weights), and consequent voter attitudes, with consequences for resulting new coalitions and policy (as captured by new welfare weights).

Finally a key aspect missing from the current model is the role of private information and communication. A substantial literature has looked at public backlash, driven by populist parties or media portrayal of IOs, where individuals come to believe that the benefits of the coordinated IO policy do not outweigh the costs of losing national autonomy over the policy-making process (Dahl, 1999; Schmidt, 2006; Hix et al., 2007; Hix, 2008; Keohane et al., 2009; Hobolt & Tilley, 2014; De Vries, 2018; Walter, 2021; Schlipphak et al., 2022; Chaudoin, 2023). There are now many examples where domestic political actors politicize international cooperation as detrimental to national sovereignty and individual welfare (De Vries et al., 2021; von Borzyskowski & Vabulas, 2024a). These influences would arise where citizens have private information about IO policies, resources, constitutional constraints or take private actions that influence demands. Future research can theoretically and empirically incorporate these aspects into the model and apply them to particular IOs or countries.

These open questions pave the way for future research and our analysis demonstrates that the structure of domestic coalitions and the allocation of welfare weights are central to understanding how international organizations function and evolve. Ultimately, our analysis reveals that citizen-driven demands shape global governance as much as institutional constraints. By bridging micro-level preferences with macro-level outcomes, this framework not only clarifies existing debates, but also paves the way for future research on the evolving role of sovereignty, representation, and legitimacy in international cooperation.