https://orcid.org/0000-0002-6821-6516
Abstract
Amidst multiple transnational crises, global governance has retaken center stage in academic and public debates. While previous generations of thinkers and citizens vigorously discussed the perennial idea of a world government, such proposals are nowadays often discarded quickly among scholars and practitioners. However, we know little about citizens’ present-day attitudes toward world government proposals. In a survey experiment on more than 42,000 citizens in 17 countries in the global South, North, East, and West, we find that the idea is only rejected by international majorities if it remains unspecified and if we weight countries equally. Specifications as democratic and/or focused on global issues like climate change significantly increase public support and lead overwhelming majorities worldwide to favor a global government. Support is even stronger in more populous, less free, less powerful, and/or less developed countries. The only exception is the United States, where no global government specification receives majoritarian public approval. Overall, our findings show significant international support for fundamental transformations of global governance, and thus indicate to activists and policymakers that relevant reform efforts can build on widespread public endorsement.
En medio de múltiples crisis transnacionales, la gobernanza global ha vuelto a ocupar un lugar importante en los debates, tanto académicos como públicos. Si bien las generaciones anteriores de pensadores y ciudadanos debatieron firmemente la idea perenne de un gobierno mundial, hoy en día los académicos y los profesionales descartan, con frecuencia, rápidamente tales propuestas. Sin embargo, tenemos poca información referente a las actitudes de los ciudadanos hacia las propuestas de un gobierno mundial. Llevamos a cabo un experimento de encuesta con más de 42.000 ciudadanos en 17 países del Sur, Norte, Este y Occidente global, y concluimos que la idea solo es rechazada por las mayorías internacionales si esta idea no es concreta (y solo cuando se ponderan los países por igual). El hecho de incluir especificaciones como la democracia y/o centrarse en temas globales como el cambio climático aumentan significativamente el apoyo público y llevan a mayorías abrumadoras en todo el mundo a favor de un gobierno mundial. Por lo general, el apoyo es más fuerte en los países más poblados, menos libres, menos poderosos y/o menos desarrollados. La única excepción es Estados Unidos, donde ninguna de estas especificaciones gubernamentales globales recibe una aprobación pública mayoritaria. En general, nuestras conclusiones muestran un importante apoyo público internacional a las transformaciones fundamentales de la gobernanza mundial y, por lo tanto, indican a los activistas y a los responsables políticos que los esfuerzos de reforma pertinentes pueden partir de un amplio respaldo público.
Face aux nombreuses crises transnationales, la gouvernance mondiale revient sur le devant de la scène dans les débats académiques et publics. Alors que les générations antérieures de penseurs et de citoyens débattaient vivement de l’idée toujours d'actualité de gouvernement mondial, de telles propositions sont de nos jours souvent rapidement rejetées par les chercheurs et les professionnels. Cependant, nous en savons peu sur l'attitude des citoyens par rapport aux propositions de gouvernement mondial. Dans une expérience de sondage sur plus de 42 000 citoyens de 17 pays du Sud, du Nord, de l'Est et de l'Ouest, nous constatons que l'idée n'est rejetée par les majorités internationales que si elle reste vague (et seulement si l'on considère les pays à parts égales). Lorsque l'on précise l'idée en qualifiant ce gouvernement mondial de démocratique et/ou de centré sur des problématiques mondiales comme le changement climatique, le soutien public augmente drastiquement et des majorités écrasantes dans le monde entier en viennent à soutenir un gouvernement mondial. Généralement, le soutien est supérieur chez les pays plus peuplés, moins libres, moins puissants et/ou moins développés. La seule exception reste les États-Unis: aucune précision concernant le gouvernement mondial ne reçoit d'approbation d'une majorité du public. Dans l'ensemble, nos résultats montrent qu'il existe un soutien international important du public vis-à-vis de transformations fondamentales de la gouvernance mondiale, et donc signalent aux militants et aux décideurs qu'ils peuvent s'appuyer sur un large soutien populaire pour promouvoir des réformes pertinentes.
Introduction
The twenty-first century has seen multiple transnational crises—from wars and mass migration to climate change and a global pandemic. Our existing international system is struggling to resolve these crises. Thus, questions of global governance often take center stage in debates among academics, practitioners, and citizens. For many years, scholars have argued for far-reaching global governance reforms to address such transnational challenges effectively and fairly (Archibugi and Held 1995; Biermann et al. 2012). Events such as Brexit, as well as the presidencies of Trump, Bolsonaro, and Duterte, have highlighted the resurgence of right-wing populism worldwide and its rejection of supranational institutions like the European Union (EU) and the World Health Organization (Norris and Inglehart 2019; Horton 2020). After the Russian invasion in 2022, Ukrainian President Zelensky publicly urged a reform of the United Nations (UN) Security Council (The Presidential Office of Ukraine 2022). Such present-day debates about global governance and its potential reform hark back to age-old discussions about how the world should be governed. Yet, much more ambitious and once prominent ideas, such as the idea of a world government, are not as fervently debated anymore (Cabrera 2010).
The idea of global government has a long history (Kissling 2005; Yunker 2011; Lu 2015). In previous centuries, advocates such as Dante Alighieri put forth the proposal of a unitary world government with a monarch at its head (Lu 2015). Such schemes for a benign global dictatorship seemed plausible in the Middle Ages (Goodin 2013, 154), and reincarnations of them were found up until the twentieth century, as evident in Hitler’s writings, for instance (Thies 2012). The idea of a world government developed with the zeitgeist, taking on a federal and democratic shape over time. In its post-World War heyday (Weiss 2009; Cabrera 2010), the proposal of a global government was championed by scientists such as Albert Einstein (1947) and Bertrand Russell (1958), philosophers like Albert Camus (Cabrera 2010, 512), suffragettes such as Rosika Schwimmer (Lloyd and Schwimmer 1942; Threlkeld 2022), and political leaders like Jawaharlal Nehru (Gilad and Freeman 2022). In scholarly circles, related discussions have reemerged and intensified since the end of the Cold War (Kahraman 2012), for example, the idea of “cosmopolitan democracy” (Archibugi and Held 1995). Such academic debates accompanied activist efforts for the establishment of institutions like a world parliament (Falk and Strauss 2000; Leinen and Bummel 2018). Some of these efforts have borne fruit, for instance, with the creation of the International Criminal Court (Struett 2008). Yet, the far-reaching proposal of a global government seems farther out of reach than it used to be.
Nowadays, many scholars explicitly distance themselves from the idea of a world government (Nicolaidis and Shaffer 2004; Benhabib 2006; Lu 2009; Keating 2012, 860), while others do so but make proposals that amount to something quite alike (Held 1995, 230; Pogge 2002, 178; Caney 2005, 266; Szombatfalvy 2010). Some suggest that the latter approach may in part be due to misguided strategic calculations (Nili 2015). However, taking arguments against world government at face value, another reason for such dismissals—besides normative reservations—is the empirical assumption that global government proposals lack public support. This assumption is shared by both advocates (Yunker 2011, 4–5, 15, 70) and opponents (Miller 2010). For the former, it demonstrates the supposed infeasibility of a world government, whereas for the latter, it contributes to the undesirability of such an institution. But is the assumption justified?
Besides assessing the validity of empirical assumptions in normative work, the question of public attitudes toward a world government also bears great practical significance as policymakers grapple with various transnational challenges whose solutions arguably depend in part on people’s support. The question of public attitudes toward world government is thus important for academics and practitioners alike. So what does existing research tell us in this regard? Scholars have studied public opinion on multilateral climate change agreements (Bechtel and Scheve 2013), present-day international organizations (IOs) (Ecker-Ehrhardt 2012; Dellmuth et al. 2022a), and potential reforms (Hahm, Hilpert, and König 2020; Ghassim, Koenig-Archibugi, and Cabrera 2022), but less so on far-reaching alternative visions for global governance such as the idea of a world government (cf. Eckhardt 1972; Ghassim 2020; Novus and Global Challenges Foundation 2020). This is the research gap that our study addresses.
Between 2017 and 2021, we conducted several rounds of survey experiments on citizens in seventeen diverse countries from around the world. In collaboration with Dynata, Qualtrics, and YouGov, we sampled respondents for national representativeness by age, gender, region, and education. We randomly split respondents into different groups who were asked to state their opinion on a range of world government proposals. For the baseline group, the proposal remained unspecified. For the treatment groups, it was specified as democratic and/or focused on global issues such as climate change or the COVID-19 pandemic. We find that the idea of a world government is rejected by an international majority (when weighting countries equally) if it remains unspecified. Specifications of a world government as democratic and/or focused on global issues significantly increase public support, producing clear majorities in favor. Attitudes are generally more positive in more populous, less free, weaker, and/or poorer countries. The United States (US) is the exception, as no global government specification receives majoritarian public endorsement there. Thus, widespread assumptions about citizens opposing world government could partly be unjustified generalizations from the American context to the world. Overall, our findings show substantial public support worldwide for fundamental transformations of global governance. Therefore, reform efforts of activists and policymakers can build on the widespread endorsement of citizens.
Concepts and Hypotheses
The term “world government” carries strong normative connotations, and we suspect that these undertones are often perceived as negative. The phrase—while somewhat void of content—evokes a mighty, all-embracing political institution, which—among others, due to the potential concentration and abuse of power (Buckinx 2012)—can seem frightening to people confronted with the idea. This characterization of the idea of world government harks back to historical notions from several centuries ago. For instance, Dante Alighieri (as cited in Lu 2015) argued for the establishment of a universal monarchy to promote human unity and world peace: “There must therefore be one person who directs and rules mankind, and he is properly called 'Monarch' or 'Emperor.' And thus it is apparent that the well-being of the world requires that there be a monarchy or empire.”
Such visions of a universal monarchy have provided the basis for critiques by various authors over the centuries. For instance, Kant (1795, 38) argued that a world government would bear the danger of turning into a global tyranny—or “soulless despotism,” as he called it. While Kant’s (1795, 24) views on the issue of a “world republic” are much more complex than often portrayed and by no means exclusively dismissive (Pogge 2009; Pinheiro Walla 2017), his statements on the threat of a global tyranny have been cited by other scholars in their rejections of a world government (Rawls 2001, 36; Slaughter 2003, 83; Nussbaum 2006, 313–4).
Goodin (2013, 154) considers such critiques to be “disingenuously inattentive” and argues that the proposal of “a unitary world government ruling over the whole world” is “simply not on the table” anymore. Nonetheless, we suspect that the majority of people around the world share such fears about the concept of world government when it is left without further specification. In this context, it has been argued that “[c]onscious or unconscious identification of world government as a form of imperial government is an important reason why the idea of world government is currently rejected by a large majority of the world’s population” (Yunker 2011, 15; emphasis added). However, no data is provided to back up this claim. Our research shows whether this empirical assertion holds. Thus, our first prediction is as follows:1
Majorities of people reject the proposal of an unspecified world government.
The world government idea appears to follow the zeitgeist: Nowadays, world government is generally envisioned as democratic. The normative justificatory basis for this is often procedural, aiming for greater “input” legitimacy of the proposed institution (Scharpf 1999). That is, advocates argue that if a global government is established, it should be democratic so that people worldwide can participate in its decision-making. Perhaps the most prominent example of such an argument is the case for “cosmopolitan democracy,” pioneered by Archibugi and Held (1995).
While the argument for cosmopolitan democracy is extensive and complex, one of its essential characteristics relates to normatively necessary extensions of democracy in the globalized age. Held (1995) argues that in the present era of globalization, where many political decisions that are directly relevant to people’s lives have shifted to the global level, there needs to be democratic control over such decisions if we want to maintain autonomy over our lives. Such democratic control could be realized, for instance, through elections to a global parliament (Falk and Strauss 2000). However, other ways of public participation are also conceivable (Scholte 2014), for example, aspects of direct democracy like in Switzerland (Kriesi and Wisler 1996) or the principle of sortition as practiced in Ancient Greece (Bouricius 2013) and more recently in citizens’ assemblies at the local, national, and global levels (Dryzek, Bächtiger, and Milewicz 2011; Global Assembly Team 2022; Goldberg and Bächtiger 2023).
Scholars have found public support for international cooperation (Bernauer, Mohrenberg, and Koubi 2020), IOs (Dellmuth, Scholte, and Tallberg 2019), and proposed global governance reforms (Ghassim, Koenig-Archibugi, and Cabrera 2022) to be driven, in part, by democratic characteristics. Therefore, we suspect that once the envisioned world government is characterized in democratic terms (as it often is), support for the idea will expand significantly. One complication here is that the concept of democracy is understood differently across the world (Kirsch and Welzel 2019), as well as having been captured and misappropriated by autocratic regimes such as the Democratic People’s Republic of Korea. In contrast, our definition and concept of democracy and a democratic world government are explicitly liberal. In particular, our hypothesis is that when the proposed global government is specified as representing people through elections or other ways of citizen participation, more people support it on average than when it is left unspecified. Our second hypothesis is thus as follows:1
An institutional specification as democratic leads more people to support the proposal of a world government compared to the unspecified version.
Similar to institutional specifications of world government, functional specifications have followed the zeitgeist as well. For instance, advocacy for world government after the Second World War (WWII) was very much focused on the newly emerged threat of nuclear weapons and characterized by the fear of a third World War that might lead to the destruction of human civilization (Russell 1958). Indeed, commentators at the time believed that a global catastrophe such as WWII should and would act as a catalyst for major global institutional changes like the establishment of a global government (Reves 1945). Such thinking is reflected in the work of contemporary scholars (Deudney 2006) who see a teleological development toward the establishment of a world state (Epps 1995; Wendt 2003). In contrast, the wave of world government advocacy since the end of the Cold War has mainly been characterized by concerns for global socioeconomic and climate justice (Soroos 1990; Pogge 2002; Caney 2005; Cabrera 2006; Biermann and Bauer 2017). The normative justificatory basis for such accounts is generally instrumental, aiming for greater “output” legitimacy for the envisioned institution (Scharpf 1999). That is, world government is justified primarily with reference to the goals and goods that it is supposed to help realize, for example, world peace, environmental protection, socioeconomic justice, or global health.
In this context, it should be noted that there is widespread agreement among global democracy and world government scholars on the principle of subsidiarity (Held 1995; Newcombe 2001; Pogge 2002), which encapsulates a common way of dividing areas of responsibility between different governance layers, for example, as it is practiced in the EU. Based on this principle, decision-making authority on any given issue should be located at the lowest possible and highest necessary layer in the governance hierarchy (cf. Føllesdal 1998). Of course, determining this is not straightforward in every instance and is always subject to negotiations. In the case of the EU, there are some policy areas (for example, trade) on which European institutions generally have the prerogative, whereas in others (for instance, education), the nation-states constituting the EU remain in charge. Moreover, within the individual nation-states, responsibility for policy areas may be delegated further down the layers of governance, as in the case of German states’ relative autonomy over the educational system. Notwithstanding different opinions on this matter in the academic literature, many theorists and practitioners concur that in case a global government is established, non-global issues should remain within the realm of lower levels of decision-making, for example, at the national level, as illustrated by Pogge’s (2002, 181) principle of a “vertical dispersal of sovereignty.”
Long-standing concerns with international peace and world poverty, as well as the preeminence of global issues like climate change in contemporary political debates, may evoke citizens’ desire to address such problems more effectively. Indeed, prior research has shown that perceived interdependence, moderated by national vulnerability (that is, the extent to which the national government is perceived as incapable of addressing a particular policy issue by itself), helps explain public support for supranational organizations (Ecker-Ehrhardt 2012). In this context, establishing a global government with wide-ranging authority and power on transnational issues may be seen as a potentially effective institutional approach. We therefore hypothesize that such functional specifications raise public support for world government compared to an unspecified version of it (Hypothesis 3).
While we were conducting this research project, the COVID-19 pandemic broke out, offering another research avenue for our project. History has shown that major calamities have served as stimuli for changes in global governance. Cases in point are the First World War, followed by the founding of the League of Nations, and the WWII, after which the UN and Bretton Woods institutions were established. The creation of the International Criminal Court (which followed the Rwandan genocide and Yugoslavian civil wars, as well as the subsequent establishment of international criminal tribunals on these conflicts) serves as a more recent example of the link between major catastrophes and watershed moments in global governance (Ljuboja 2009; Chlevickaitė, Holá, and Bijleveld 2020). Consequently, scholars and activists have argued that the establishment of a world government may require a global-scale disaster (cf. Reves 1945; Wendt 2003; Deudney 2006; Leinen and Bummel 2018, 394–7). In this vein, commentators have suggested global integration as an institutional response to fighting the coronavirus (Bummel 2020). As the COVID-19 pandemic unfolded, we expected—in line with our broader hypothesis on functional specifications—that an acute global issue such as this, affecting the lives of people everywhere, would make citizens more supportive of the idea of a global government compared to the unspecified vision of such an institution. Our next hypothesis is thus as follows:1
Functional specifications of world government as focused on global issues cause more people to support the proposal compared to the unspecified version.
While different theorists focus on diverging justificatory strategies for their accounts, most proposals for world government indeed feature both institutional and functional specifications (DuFord 2016). For instance, in his model of “cosmopolitan democracy,” which is largely built on procedural grounds, Held (1995, 235) envisions embedding the principle of subsidiarity, such that governmental institutions at each level—from the global to the subnational—should only be charged with those issues that are properly handled at the respective layer and cannot be dealt with at lower levels of governance. Similarly, theorists who rely primarily on instrumental justifications in their case for a world government often specify their proposals institutionally as democratic. For instance, while Cabrera’s (2006) cosmopolitan argument for a world state is principally concerned with achieving global justice, he insists that such an institution should be democratic.
There are also some advocates of world government, however, who do not call for such an institution to have democratic structures and be governed by popular participation. Deudney’s (2006, chap. 9) argument for “federal-republican nuclear one worldism” is a case in point. He argues for a limited mutual restraint union at the global level focused on managing the most dangerous security threats to mankind. Deudney (2006, 249–52) explicitly juxtaposes this idea with notions of “classical nuclear one worldism,” which involved the idea of a world state that was often envisioned as democratic, for example, by means of a classic separation of powers (Clark and Sohn 1966). Hence, given the insistence of some authors that their visions for world government would not necessarily be democratic, it is worth distinguishing between institutional and functional specifications of world government advocacy (as we do), even though in many cases for world government these justifications are combined. When such specifications overlap, we conjecture—in line with Hypotheses 2 and 3—that they increase people's support for world government compared to the baseline scenario of an unspecified world government. The same reasoning applies to the combination of a democratic institutional specification with a functional focus on the COVID-19 pandemic. Our next hypothesis is therefore as follows:1
An institutional and functional specification as democratic and focused on global issues makes people more likely to support the proposal of a world government compared to the unspecified version.
Now, let us return to the recurring but unsubstantiated empirical claim that people worldwide would not support the establishment of a world government. We also suspect that most people would reject the proposal of an unspecified specter of world government (Hypothesis 1). Yet, we conjecture that majorities worldwide would support the establishment of a world government once it is specified as normative theorists often do, that is, in institutional terms as democratic and functionally concentrated on global issues like climate change, world poverty, and international peace. The same reasoning applies to the combination of a democratic setup with a functional focus on the COVID-19 pandemic. These hypotheses are based on the assumption that the perceived inadequacy of the existing international system may motivate many people to support very different and much stronger supranational organizations than those that currently exist. Moreover, the results of substantively related surveys suggest that proposals for strong supranational organizations may indeed command majoritarian public support (Ghassim 2020; Novus and Global Challenges Foundation 2020). Our hypothesis in this context is thus as follows:1
Majorities of people support the proposal of a world government when it is institutionally and functionally specified as democratic and focused on global issues.
Besides the hypothesized treatment effects and majority opinions above, we expected that public attitudes toward the proposal of a world government could be explained, in part, by country-level factors such as population size, economic development, regime type, and material power, and that such factors could condition how different specifications influence public attitudes toward world government. While expectations in this respect can be separated in theory, they may be hard to distinguish analytically as the factors tend to correlate. For example, population size and development may be correlated with national power, while power and development could be associated with regime type (as developed and dominant Western powers tend to be more democratic). Overlaps such as these are also evident in our theoretical discussion below. Thus, while acknowledging that the factors cannot be neatly separated, let us now turn to developing hypotheses on these country-level factors independently. For each of them, associations with public attitudes toward global government could conceivably work in both directions.
First, citizens of more populous countries may be more supportive of a world government if they expect to have a greater say on the global stage as a result. Our world government specifications conceivably give more populous nations a much bigger say in global affairs. Assuming people believe that their compatriots’ attitudes are generally more aligned with their own than the attitudes of foreign nationals, which is not necessarily the case (Hale and Koenig-Archibugi 2019), we would expect that citizens of more populous countries are more supportive of a world government than citizens of less populous countries. This may be especially true for our specifications of a world government as democratic (that is, the institutionally and fully specified proposals), as they conceivably include a world parliament which may be based on principles such as population-based proportional representation.
Alternatively, citizens of less populous countries may be more supportive of a world government if they think that their voices would be better heard than in the current international system. At present, geopolitical power is arguably skewed toward countries with greater populations, given that the size of a nation is an important determinant of its military and economic might (Singer 1988). Citizens of less populous countries may perceive transformative changes in global governance, such as the establishment of a world government, as moves toward a global order in which population size plays a smaller role in determining their status in the world. Our diverging expectations in this regard may be summarized as follows:
Citizens of more populous countries tend to be more supportive of a world government.
Citizens of less populous countries tend to be more supportive of a world government.
Second, citizens of more economically developed countries may be more supportive of a global government if they believe that their currently privileged status would be amplified in such an alternative world order. In the present international system, a country’s level of development is an important determinant of its global status. This is true for bilateral relations between countries, but is also manifested in the institutional structures of organizations such as the International Monetary Fund where vote shares are determined by countries’ financial contributions which, in turn, are linked to (historical) national economic productivity (Bryant 2008). Strengthening global governance, for instance, through the creation of a world government, may amplify the privileged positions of more developed countries and allow them to benefit further, thus mustering greater support from their citizens.
Alternatively, citizens of more developed states may oppose a world government if they fear that it would result in global wealth redistribution, which in turn may be a primary motive for citizens of less developed countries to be more supportive of such proposals. Indeed, normative scholars have claimed that concerns about global redistribution are a key reason why citizens of more developed countries oppose wide-ranging global governance reform (Scheuerman 2011, 164–5). If this is the case, then we should expect the democratic specification to create a greater cleavage between more and less developed countries than the global issues specifications, since respondents may be more likely to expect global redistribution in the former if they associate it with institutions such as a world parliament where global redistribution proponents could hold a majority.
A classic argument in the empirical literature is that “cognitive mobilization” induces greater support for the EU (Inglehart 1970) and other IOs (Ecker-Ehrhardt 2012, 26), although this association is not always significant (Dellmuth et al. 2022b, 296). Knowledge of IOs, in turn, is associated with economic aspects such as a country’s development status (Torgler 2008, 6) and income inequality (Dellmuth 2016). Thus, factors like political awareness, education, and interest may act as mediators in the hypothesized relationships between a country’s development and its citizens’ attitudes toward a world government. Bearing such potential mediating factors in mind, we focus here on the direct association between economic development and attitudes toward world government. Our contrasting expectations in this regard may be summarized as follows:
Citizens of more developed countries tend to be more supportive of a world government.
Citizens of less developed countries tend to be more supportive of a world government.
Third, citizens of freer and more democratic countries may be more supportive of a world government if they think that such an institution would help spread the norms of democracy and freedom. Our conception of a global issues-focused and/or democratic world government carries the potential of widening the scope and applicability of liberal norms, as reflected in many such proposals over the centuries (Cabrera 2010). If citizens believe such a scenario to be realistic, they may be more likely to support the creation of a world government than citizens of less free and democratic countries.
Alternatively, citizens of freer and more democratic countries may be less supportive if they expect their privileges to be threatened by a world government. It has been argued that even a world government intended to strengthen liberal democratic norms could end up undermining them, for instance, due to global heterogeneity resulting in civil strife or rule by illiberal global majorities (cf. Kant 1795; Pogge 2009; Pinheiro Walla 2017). A related fear is that citizens in nondemocratic countries may ultimately constitute worldwide majorities and that widespread illiberal conceptions of democracy (Kirsch and Welzel 2019) would shape the politics of any supposedly democratic global government.
Conversely, citizens of less democratic and free countries may be more likely to support a world government as conceived here. This could be due to the belief that the creation of such an institution would help them obtain greater privileges of freedom and democracy, which they are deprived of in the current international system, and/or if such a global government would pressure their domestic regime to democratize (assuming that they desire such a democratization). All of these arguments are especially pertinent with respect to world government proposals that contain democratic specifications, as those are most likely to materialize the fears and hopes of people in more or less free countries. The corresponding contradictory expectations may be summarized as follows:
Citizens of freer countries are more likely to support a world government.
Citizens of less free countries are more likely to support a world government.
Fourth, a country’s power may help explain public attitudes toward world government proposals. As one of the central concepts in international relations, power has long been much debated in the academic literature (Holsti 1964). It is commonly conceptualized as implying certain material capabilities, such as military personnel and equipment, population size, and economic strength (Singer 1988). As a decisive factor in global politics, public opinion toward world government may be influenced by countries’ power, and—once again—expectations of such an association may legitimately go in both directions.
Citizens of more powerful countries may be more supportive of a global government if they believe that their power position would be magnified in such an alternative world order which may seem likely given the privileges of more powerful countries in present-day global governance. Moreover, as noted above, one of the long-lasting fears of world government is that of global dictatorship (Kant 1795). While this would constitute a nightmarish scenario for many, citizens of the most powerful countries may expect even greater privileges as a result and thus support the idea. Alternatively, citizens of more powerful countries may be less supportive of a world government if they think that such an institution could cause them to lose their privileged status (cf. Scheuerman 2011).
Similarly, citizens of weaker countries may be more supportive of a world government if they believe that the creation of such an institution would improve their status in the global order compared to the current international system in which their countries’ standing disadvantages them in relation to citizens of more powerful countries. For them, expected benefits of global integration may range from issues such as visa privileges to benefits of governmental protection in crisis situations (which are arguably skewed toward more powerful countries at present). Related arguments were made in the context of the EU’s establishment (Grieco 1995). Since these considerations largely relate to the transfer of sovereignty over certain policy issues from the national to the global level, we may expect them to be especially pertinent to the functional world government specifications rather than the institutional ones. That is, public attitudes of citizens in more or less powerful countries may be more associated with the world government’s focus on global issues than whether it is specified as democratic.
Some prior research argues that citizens are not necessarily more confident in the IOs if their country is better represented (Dellmuth and Tallberg 2015), albeit relying on rather strong assumptions that citizens are aware of their country’s (even temporary) membership status in the UN Security Council and the involvement level of domestic NGOs in the UN. Focusing on the United States, Brutger and Clark (2023) find that partisanship moderates how citizens weigh elite cues about financial burdens against the influence over IOs that funding yields. However, there is a lack of research on the potential association between country power and individual attitudes toward the idea of a world government. Our last set of diverging expectations can be summarized as follows:
Citizens of more powerful countries tend to be more supportive of a world government.
Citizens of less powerful countries tend to be more supportive of a world government.
Data and Methods
In our survey experiments, we used “direct treatments” (Mutz 2011, chap. 3). That is, for our randomly selected experimental groups, we systematically modified the prompt they were presented, compared to the baseline group, who saw the unspecified proposal. Due to the completely random allocation of respondents into experimental conditions, the treatment effects are simply the differences-in-means between the different groups’ responses. As conceptualized above, the different versions of the global government proposal were: an unspecified world government; a democratic world government allowing for citizen participation; a world government focused on global issues such as international peace, world poverty, and climate change, or—as a separate treatment group in our last survey round—the coronavirus pandemic; and a world government combining these institutional and functional specifications (namely, being democratic and focusing on global challenges). We asked our survey respondents to what extent they endorse or reject these ideas.
Our survey experiment’s baseline group was presented with a question on the extent to which they support or oppose the establishment of an unspecified world government—see online appendix, table AT1. Our operationalization left the idea of a world government intentionally vague so that respondents could project their hopes and fears into it, depending on whether the term “world government” evoked positive or negative feelings in them.
Respondents in the institutional specification treatment group were asked about their support or opposition with regard to a “democratic” world government for which it was specified that “people worldwide would be represented through free and fair elections or other ways of citizen participation” (see online appendix, table AT1). Note that we employed an explicitly minimalist definition of democracy by reducing the concept to input aspects rather than including output considerations (Dahl 1989; Scharpf 1999). In our age of representative parliamentary democracy, free and fair elections are generally considered the standard way of enabling popular participation. However, we did allow for considering means of citizen input other than elections, for example, direct democracy or sortition, by including the intentionally broad phrase “other ways of citizen participation.” We thereby focus respondents’ attention on the general principle of democracy rather than democracy through the specific means of elections. With our operationalization, we thus anticipated diverse understandings of democracy and aimed to invoke a procedural conception in respondents’ minds when stating their attitude toward a democratic world government.
In our first functional specification condition, respondents were exposed to the proposal of a world government with “the right and the power to deal with global issues like climate change, world poverty, and international peace, while national governments would maintain control over issues that are not global” (see online appendix, table AT1). The first thing to note is our specification of the envisioned world government as having the right and the power to address global issues. This phrase implies both legal authority and material capacity to take and enforce decisions on certain global policy issues. Furthermore, our treatment focuses on three of the most frequently cited global issue areas in world government advocacy, namely global environmental, security-related, and socioeconomic challenges. Additionally, we paraphrased the principle of subsidiarity (Føllesdal 1998), which generally forms part of present-day world government proposals (Newcombe 2001).
Our second functional specification condition, which only featured in our third survey round in 2021, inquired about a world government specifically to address the COVID-19 pandemic. We asked respondents about their opposition or support for the proposal of a world government with “the right and the power to deal with global issues like the current coronavirus pandemic, while national governments would maintain control over issues that are not global” (see online appendix, table AT1). The justifications for our wording are the same as for the first functional specification condition above. The only difference here was that the functional focus of the proposed world government was exemplified by the contemporary coronavirus pandemic.
Our fully specified world government proposal combined the characteristics of the institutional and first functional specification groups above, that is, “a world government which should be democratic in that people worldwide would be represented through free and fair elections or other ways of citizen participation, and which should have the right and the power to deal with global issues like climate change, world poverty, and international peace; while national governments would maintain control over issues that are not global” (see online appendix, table AT1). The second fully specified proposal combined the idea of a democratic world government with a functional focus on COVID-19. Online appendix, table AT1 presents further details on our survey design and implementation, as well as the wording of our different experimental conditions and dependent variable questions.2
Participants in all experimental groups were offered six-point scales for their responses: “strongly oppose,” “oppose,” “somewhat oppose,” “somewhat support,” “support,” and “strongly support” (see online appendix, table AT1). These six points allowed our respondents in different countries to issue opinions of sufficient detail for our purposes. We intentionally did not offer a middle option—neither numerically nor through a neutral statement like “neither agree nor disagree”—because the meaning of such a response would not seem intuitively plausible or easily distinguishable from a nonresponse (Sturgis, Roberts, and Smith 2014). We also did not offer an “I don’t know” option in our second and third main rounds, thereby following the advice of various scholars (Krosnick et al. 2002), who argue that the inclusion of such an option as an explicit answer choice induces “satisficing” (Krosnick, Narayan, and Smith 1996) among respondents and discourages meaningful responses that would be provided in its absence. This concern was confirmed in our pilot study via Amazon Mechanical Turk in the United States in 2017, where we observed that substantial numbers of respondents who selected the no-opinion choice actually had substantive opinions as expressed in their comments. We thus decided to remove the “I don’t know” option from our answer choice scale in our main survey rounds.3 However, in order to prevent forcing genuinely unopinionated respondents into a “non-attitude” (Converse 1970), we allowed respondents to skip questions and thereby refrain from providing a response altogether. Incidentally, we still received disproportionately many item nonresponses in our US survey across all experimental groups (see table 1), indicating American respondents’ divergent behavior and attitudes on the question of a world government, as described further below, and suggesting that “don’t know” responses on the question of a world government may not have been an issue in most countries, even if we had offered the option explicitly.
Sample size and item nonresponses by country
| Country | Sample size | Item nonresponses | Percentage of nonresponses |
|---|---|---|---|
| Argentina | 1,035 | 1 | 0.1 |
| Australia | 3,122 | 9 | 0.3 |
| Canada | 3,100 | 7 | 0.2 |
| China | 1,263 | 0 | 0 |
| Colombia | 3,275 | 5 | 0.2 |
| Egypt | 3,100 | 30 | 1.0 |
| France | 3,609 | 16 | 0.4 |
| Hungary | 3,100 | 6 | 0.2 |
| India | 1,034 | 0 | 0 |
| Indonesia | 3,231 | 9 | 0.3 |
| Kenya | 3,250 | 11 | 0.3 |
| Russia | 2,578 | 0 | 0 |
| South Korea | 3,100 | 8 | 0.3 |
| Spain | 1,017 | 1 | 0.1 |
| Turkey | 3,432 | 9 | 0.3 |
| UK | 1,659 | 367 | 22.1 |
| USA | 1,063 | 328 | 30.9 |
| Total | 41,968 | 807 | 1.9 |
| Country | Sample size | Item nonresponses | Percentage of nonresponses |
|---|---|---|---|
| Argentina | 1,035 | 1 | 0.1 |
| Australia | 3,122 | 9 | 0.3 |
| Canada | 3,100 | 7 | 0.2 |
| China | 1,263 | 0 | 0 |
| Colombia | 3,275 | 5 | 0.2 |
| Egypt | 3,100 | 30 | 1.0 |
| France | 3,609 | 16 | 0.4 |
| Hungary | 3,100 | 6 | 0.2 |
| India | 1,034 | 0 | 0 |
| Indonesia | 3,231 | 9 | 0.3 |
| Kenya | 3,250 | 11 | 0.3 |
| Russia | 2,578 | 0 | 0 |
| South Korea | 3,100 | 8 | 0.3 |
| Spain | 1,017 | 1 | 0.1 |
| Turkey | 3,432 | 9 | 0.3 |
| UK | 1,659 | 367 | 22.1 |
| USA | 1,063 | 328 | 30.9 |
| Total | 41,968 | 807 | 1.9 |
Notes: The table shows item nonresponse numbers and their percentage of the total sample size on any of our world government questions across the different experimental groups, along with the total sample size (including item nonresponses) in each survey country.
Sample size and item nonresponses by country
| Country | Sample size | Item nonresponses | Percentage of nonresponses |
|---|---|---|---|
| Argentina | 1,035 | 1 | 0.1 |
| Australia | 3,122 | 9 | 0.3 |
| Canada | 3,100 | 7 | 0.2 |
| China | 1,263 | 0 | 0 |
| Colombia | 3,275 | 5 | 0.2 |
| Egypt | 3,100 | 30 | 1.0 |
| France | 3,609 | 16 | 0.4 |
| Hungary | 3,100 | 6 | 0.2 |
| India | 1,034 | 0 | 0 |
| Indonesia | 3,231 | 9 | 0.3 |
| Kenya | 3,250 | 11 | 0.3 |
| Russia | 2,578 | 0 | 0 |
| South Korea | 3,100 | 8 | 0.3 |
| Spain | 1,017 | 1 | 0.1 |
| Turkey | 3,432 | 9 | 0.3 |
| UK | 1,659 | 367 | 22.1 |
| USA | 1,063 | 328 | 30.9 |
| Total | 41,968 | 807 | 1.9 |
| Country | Sample size | Item nonresponses | Percentage of nonresponses |
|---|---|---|---|
| Argentina | 1,035 | 1 | 0.1 |
| Australia | 3,122 | 9 | 0.3 |
| Canada | 3,100 | 7 | 0.2 |
| China | 1,263 | 0 | 0 |
| Colombia | 3,275 | 5 | 0.2 |
| Egypt | 3,100 | 30 | 1.0 |
| France | 3,609 | 16 | 0.4 |
| Hungary | 3,100 | 6 | 0.2 |
| India | 1,034 | 0 | 0 |
| Indonesia | 3,231 | 9 | 0.3 |
| Kenya | 3,250 | 11 | 0.3 |
| Russia | 2,578 | 0 | 0 |
| South Korea | 3,100 | 8 | 0.3 |
| Spain | 1,017 | 1 | 0.1 |
| Turkey | 3,432 | 9 | 0.3 |
| UK | 1,659 | 367 | 22.1 |
| USA | 1,063 | 328 | 30.9 |
| Total | 41,968 | 807 | 1.9 |
Notes: The table shows item nonresponse numbers and their percentage of the total sample size on any of our world government questions across the different experimental groups, along with the total sample size (including item nonresponses) in each survey country.
Within the constraints of online survey research, which leans toward more developed countries and regions with greater internet access penetration, we selected our survey countries to allow for diversity with respect to the different factors theorized above (see table 2). After our pilot survey in 2017, we conducted our first main survey round on a representative sample of the British population (excluding Northern Ireland) in March 2018 in collaboration with YouGov. Between March and June 2019, we fielded a multinational survey experiment on nationwide samples of citizens in Argentina, China, India, Russia, Spain, and the United States in collaboration with Dynata. From May to October 2021, we conducted another international survey, this time including ten countries all over the world—Australia, Canada, Colombia, Egypt, France, Hungary, Indonesia, Kenya, South Korea, and Turkey—in collaboration with Qualtrics and its survey partners (Cint, Dynata, Lucid, and Toluna). These companies’ online survey services are regularly used for academic purposes, including the recruitment of respondents (Peer et al. 2022). Individuals sign up to these databases in order to complete surveys for compensations that are set by the companies in line with standard rates in online survey research in the respective survey countries (e.g., Lucid 2023). The participation of respondents in our survey was completely voluntary and anonymous.2
Characteristics of survey countries
| Country | Population | Income | Freedom | Power |
|---|---|---|---|---|
| Argentina | 45 | 8,442 | 84 (free) | 0.005 |
| Australia | 25 | 51,812 | 97 (free) | 0.007 |
| Canada | 38 | 43,258 | 98 (free) | 0.008 |
| China | 1,439 | 10,500 | 11 (not free) | 0.231 |
| Colombia | 51 | 5,333 | 65 (partly free) | 0.007 |
| Egypt | 102 | 3,548 | 18 (not free) | 0.010 |
| France | 65 | 39,030 | 90 (free) | 0.013 |
| Hungary | 10 | 15,899 | 69 (partly free) | 0.001 |
| India | 1,380 | 1,901 | 67 (partly free) | 0.087 |
| Indonesia | 274 | 3,870 | 59 (partly free) | 0.016 |
| Kenya | 54 | 1,838 | 48 (partly free) | 0.002 |
| Russia | 146 | 10,127 | 20 (not free) | 0.036 |
| South Korea | 51 | 31,489 | 83 (free) | 0.023 |
| Spain | 47 | 27,063 | 94 (free) | 0.008 |
| Turkey | 84 | 8,538 | 32 (not free) | 0.015 |
| UK | 68 | 40,285 | 94 (free) | 0.013 |
| USA | 331 | 63,544 | 86 (free) | 0.133 |
| Country | Population | Income | Freedom | Power |
|---|---|---|---|---|
| Argentina | 45 | 8,442 | 84 (free) | 0.005 |
| Australia | 25 | 51,812 | 97 (free) | 0.007 |
| Canada | 38 | 43,258 | 98 (free) | 0.008 |
| China | 1,439 | 10,500 | 11 (not free) | 0.231 |
| Colombia | 51 | 5,333 | 65 (partly free) | 0.007 |
| Egypt | 102 | 3,548 | 18 (not free) | 0.010 |
| France | 65 | 39,030 | 90 (free) | 0.013 |
| Hungary | 10 | 15,899 | 69 (partly free) | 0.001 |
| India | 1,380 | 1,901 | 67 (partly free) | 0.087 |
| Indonesia | 274 | 3,870 | 59 (partly free) | 0.016 |
| Kenya | 54 | 1,838 | 48 (partly free) | 0.002 |
| Russia | 146 | 10,127 | 20 (not free) | 0.036 |
| South Korea | 51 | 31,489 | 83 (free) | 0.023 |
| Spain | 47 | 27,063 | 94 (free) | 0.008 |
| Turkey | 84 | 8,538 | 32 (not free) | 0.015 |
| UK | 68 | 40,285 | 94 (free) | 0.013 |
| USA | 331 | 63,544 | 86 (free) | 0.133 |
Notes: Population figures are provided in millions of people and rounded to the nearest million (United Nations 2022). Average national income is calculated as GDP per capita in US dollars (The World Bank Group 2021). Freedom House (2021) assesses countries’ political rights and civil liberties annually, and we used the sum of these scores for the country-year in question, on a range from 0 to 100. The National Material Capabilities (NMC) score evaluates countries’ power and theoretically ranges from 0 to 1, weighting six factors equally: military personnel, military expenditures, total population, urban population, iron and steel production, as well as primary energy consumption (Singer 1988; Correlates of War 2021).
Characteristics of survey countries
| Country | Population | Income | Freedom | Power |
|---|---|---|---|---|
| Argentina | 45 | 8,442 | 84 (free) | 0.005 |
| Australia | 25 | 51,812 | 97 (free) | 0.007 |
| Canada | 38 | 43,258 | 98 (free) | 0.008 |
| China | 1,439 | 10,500 | 11 (not free) | 0.231 |
| Colombia | 51 | 5,333 | 65 (partly free) | 0.007 |
| Egypt | 102 | 3,548 | 18 (not free) | 0.010 |
| France | 65 | 39,030 | 90 (free) | 0.013 |
| Hungary | 10 | 15,899 | 69 (partly free) | 0.001 |
| India | 1,380 | 1,901 | 67 (partly free) | 0.087 |
| Indonesia | 274 | 3,870 | 59 (partly free) | 0.016 |
| Kenya | 54 | 1,838 | 48 (partly free) | 0.002 |
| Russia | 146 | 10,127 | 20 (not free) | 0.036 |
| South Korea | 51 | 31,489 | 83 (free) | 0.023 |
| Spain | 47 | 27,063 | 94 (free) | 0.008 |
| Turkey | 84 | 8,538 | 32 (not free) | 0.015 |
| UK | 68 | 40,285 | 94 (free) | 0.013 |
| USA | 331 | 63,544 | 86 (free) | 0.133 |
| Country | Population | Income | Freedom | Power |
|---|---|---|---|---|
| Argentina | 45 | 8,442 | 84 (free) | 0.005 |
| Australia | 25 | 51,812 | 97 (free) | 0.007 |
| Canada | 38 | 43,258 | 98 (free) | 0.008 |
| China | 1,439 | 10,500 | 11 (not free) | 0.231 |
| Colombia | 51 | 5,333 | 65 (partly free) | 0.007 |
| Egypt | 102 | 3,548 | 18 (not free) | 0.010 |
| France | 65 | 39,030 | 90 (free) | 0.013 |
| Hungary | 10 | 15,899 | 69 (partly free) | 0.001 |
| India | 1,380 | 1,901 | 67 (partly free) | 0.087 |
| Indonesia | 274 | 3,870 | 59 (partly free) | 0.016 |
| Kenya | 54 | 1,838 | 48 (partly free) | 0.002 |
| Russia | 146 | 10,127 | 20 (not free) | 0.036 |
| South Korea | 51 | 31,489 | 83 (free) | 0.023 |
| Spain | 47 | 27,063 | 94 (free) | 0.008 |
| Turkey | 84 | 8,538 | 32 (not free) | 0.015 |
| UK | 68 | 40,285 | 94 (free) | 0.013 |
| USA | 331 | 63,544 | 86 (free) | 0.133 |
Notes: Population figures are provided in millions of people and rounded to the nearest million (United Nations 2022). Average national income is calculated as GDP per capita in US dollars (The World Bank Group 2021). Freedom House (2021) assesses countries’ political rights and civil liberties annually, and we used the sum of these scores for the country-year in question, on a range from 0 to 100. The National Material Capabilities (NMC) score evaluates countries’ power and theoretically ranges from 0 to 1, weighting six factors equally: military personnel, military expenditures, total population, urban population, iron and steel production, as well as primary energy consumption (Singer 1988; Correlates of War 2021).
Our sample of countries represents 54 percent of the contemporary global population (United Nations 2022). We used census-based sampling quotas for various demographics (including gender, age, region, and—in our first and third main rounds—education) to make our samples as nationally representative as possible (see online appendix 1.2). The final sample sizes are presented in table 1, amounting to an aggregate sample of 42,469 respondents (including 501 respondents on Amazon Mechanical Turk in our pilot survey).
We used the randomization function on Qualtrics to split each country sample into experimental conditions of approximately equal sizes. We employed entropy balancing (Hainmueller 2012) to reweight the raw sample sizes of complete responses to the proportions of the target quotas by gender, age, region, and education (see online appendix 1.2). In the case of our UK survey, YouGov implemented both the randomization and created the weights to ensure the national representativeness of our results. For our cross-country analyses using equal weights, each country sample is weighted down to 1,000, whereas in our cross-country analyses using population weights, each country sample is weighted proportionally to its population size (see table 2).
The main treatment effects are the differences-in-means between the different experimental conditions. Student’s (1908) t-tests are employed to establish the statistical significance of differences-in-means using conventional levels (starting with p < 0.1 as weak significance). We employ multivariate regressions—using ordinary least squares (OLS), ordered logistic (o-logit) regressions, and multilevel mixed-effects analysis—to investigate potential associations between country-level factors and attitudes toward different world government proposals.4
Results and Discussion
Our study shows that empirical assumptions about a lack of public support for world government are largely unfounded, as they only tend to apply to an unspecified notion of a global government when weighting countries equally, but not to common present-day specifications of such an institution as democratic and focused on transnational issues. In fact, when specified as such, overwhelming majorities worldwide endorse the idea. Investigating potential explanatory factors at the country-level, we find that citizens of more populous, less developed, less free, and/or less powerful countries are generally even more supportive, all else equal. This indicates that people may view the idea of a democratic and functionally focused global government as a way of overcoming inequalities in development and power, advancing their nation’s preferences in world politics, and acquiring greater freedoms. In the following, we present and discuss these results in greater detail.
Confirming our expectation (Hypothesis 1), the proposal of an unspecified world government is rejected by a majority of 52 percent when weighting each country in our sample equally (see figure 1). However, when weighting each country proportionally based on its population, the unspecified world government proposal is endorsed by a majority of 58 percent (see figure 2). Thus, our surveys indicate that the empirical assumption of a world government being rejected by the popular majority (Hypothesis 1) does not even fully hold in the case of an unspecified proposal. The empirical basis for this claim becomes even more questionable once the global government proposal is institutionally and functionally specified.
Response proportions by condition across countries, using equal weights. Notes: Potential deviations from 100 percent in each row are due to rounding. The experimental conditions referring to COVID-19 only featured in the third survey round, while the experimental condition “Global issues” was not included in this form in the 2018 UK survey (see online appendix, table AT1).
Response proportions by condition, across all countries, using population weights. Notes: See notes in figure 1.
As predicted in Hypotheses 2–4, support for world government in each treatment condition is significantly higher than in the base condition of an unspecified world government, using both equal and population weights—see figure 3, online appendix, figure AF1 and tables AT43–AT46—indicating that both input and output considerations motivate public support. This is confirmed by our o-logit robustness checks (see online appendix, tables AT44 and AT46). The proposal of a democratic world government is endorsed by 67 and 71 percent when using equal and population weights, respectively (see figures 1 and 2). Support for a world government focusing on a broad set of global issues is at 66 and 69 percent using equal and population weights, respectively (see figures 1 and 2). A world government focusing on COVID-19 is supported by 64 and 69 percent, using equal and population weights, respectively (see figures 1 and 2).
Mean effects by treatment condition, across countries, using equal weights. Notes: The response scale ranges from 1 (strongly oppose) to 6 (strongly support), thus amounting to five units. Here, we show the effect sizes of each world government specification compared to the unspecified version, which has been normalized to zero in this plot. The lines indicate 95 percent confidence intervals around the point estimate. All differences-in-means between the specified conditions and the unspecified condition are highly statistically significant at conventional levels (p < 0.001). In the models, the dependent variables were regressed on a binary variable indicating unspecified (0) or specified condition (1) without including any control variables since the covariates are distributed equally due to treatment randomization. The experimental conditions referring to COVID-19 only featured in the third survey round, while the experimental condition “Global issues” was not included in this form in the 2018 UK survey (see online appendix, table AT1). Detailed results of the OLS regressions underlying this plot are presented in the online appendix, table AT43. An o-logit robustness check is provided in the online appendix, table AT44.
Hypothesis 5 regarding majority opinions in favor of institutionally and functionally specified world government proposals is also corroborated. The idea of a democratic and global-issues-focused world government is endorsed by 69 and 72 percent, using equal and population weights, respectively (see figures 1 and 2), whereas 72 and 74 percent of citizens (using equal and population weights, respectively) endorse the idea of a democratic global government focusing on COVID-19—the highest ratings among our specifications (see figures 1 and 2). While our immediate interest was in the causal effects that a COVID-19 focus has on public support for a world government, our findings are also informative with regard to the possible influence of future pandemics on public attitudes in this regard.
Our results are broadly in line with recent survey research that finds international public support for a global democracy including a global government and parliament (Ghassim 2020), as well as a supranational organization mandated to make binding decisions on global issues (Global Challenges Foundation 2014, 2017; ComRes and Global Challenges Foundation 2018; Novus and Global Challenges Foundation 2020).
For our investigation of potential explanatory factors at the country-level, we conducted multivariate OLS regression analyses, as well as o-logit and multilevel analyses as robustness checks, including our four suspected country-level factors as independent variables, using equal and population weights (see online appendix, tables AT47–AT52). The results show relatively consistently that world government support is positively associated with population size and negatively associated with material power, political freedom, and economic development. These results are illustrated in figures 4 and 5, using equal and population weights for the democratic and global issues specification. Online appendix 2.2 presents the response proportions by country types for all other experimental conditions separately, using equal and population weights.
Full specification results, across countries, equal weights, by country groups. Notes: Potential deviations from 100 percent in each row are due to rounding. The data underlying this plot comes from all three main survey rounds (see online appendix, table A1). In our binary categorizations, we aimed for substantively reasonable thresholds leading to approximately equal numbers of countries in both categories. For population size, we used 75 million as the cutoff between medium-sized and large countries. For economic development, we used 20,000 US dollars in GDP per capita as the threshold between richer and poorer countries. For regime type, we grouped partly and not-free countries into one category. For material power, we used NMC = 0.011 as the dividing line between weaker and stronger countries. However, note that these divides are mainly for illustrative purposes. Our results reported here are backed up by multivariate regression analyses that treat each variable as continuous (see table 2).
Full specification results, across countries, population weights, by country groups. Notes: See notes in figure 4.
Population size is positively associated with world government support across all experimental conditions using equal and population weights, except for the two COVID-19 treatments where the associations are statistically insignificant (see online appendix, tables AT47 and AT50), albeit clearly positive in the cross-tabulations without control variables (see online appendix, figures AF8 and AF9). These findings are largely corroborated by the o-logit robustness checks and multilevel analysis (see online appendix, tables AT48, AT49, AT51, and AT52). Citizens of more populous countries tend to be more likely to support the idea of a global government—in line with Hypothesis 6a. Moreover, when comparing the democratic and global issues specifications to the unspecified condition (see online appendix, figures AF2–AF7), it appears that this difference may be more driven by the democratic than the functional specification. This may be due to the expectation that citizens of more populous countries would benefit from such an alternative global order due to their nation’s relatively large size. Indeed, a democratic world government may benefit larger nations, assuming procedures in which population size plays a role (for example, weighted voting or direct elections to a global parliament with proportional representation) and assuming that preferences within countries are relatively homogenous compared to preferences across countries, which is not necessarily true (Hale and Koenig-Archibugi 2019).
Economic development has a generally negative association with world government support (in line with Hypothesis 7b), albeit rather consistently only when using population weights. Exceptions here are again the two COVID-19 conditions where the associations turn out insignificant, and the unspecified condition where the associations turn out (weakly) positive using both equal and population weights (see online appendix, tables AT47–AT52). However, the bivariate cross-tabulations without control variables corroborate the generally negative correlation between development level and global government support (see figures 4–5 and online appendix 2.2). The results are largely corroborated by the robustness checks using o-logits and multilevel analysis (see online appendix, tables AT48, AT49, AT51, and AT52). Overall, the results indicate that citizens of richer countries tend to be less supportive of a world government. Moreover, figures AF4–AF7 in the online appendix suggest that the democratic world government specification is linked to greater cleavages between richer and poorer countries than the global issues specification. This may be due to fears of global wealth redistribution in such an alternative world order. Indeed, that is why scholars have claimed that the idea of a world government is rejected in the developed world (Scheuerman 2011, 164–5). While our study shows that such claims are generally not true with respect to majorities in developed countries (see figures 4–5 and online appendix 2.2), this analysis indicates that development is nonetheless negatively associated with world government support.
Domestic freedoms have a generally negative association with world government support as well, in line with Hypothesis 8b. For the main analysis and robustness checks, this is consistently true for the unspecified proposal and the two COVID-19 treatments, whereas the democracy and global issues conditions yield somewhat inconsistent results, using equal and population weights (see online appendix, tables AT47–AT52). Moreover, figures AF4–AF7 in the online appendix suggest that the democratic specification is more strongly associated than the functional specification with the divergence between public support in countries with greater versus lesser freedoms. We may interpret the generally negative association as evidence for the arguments that citizens of countries with fewer freedoms may hope to gain these rights in such an alternative global order, while citizens of countries with more freedoms tend to be more likely to fear losing such privileges if a world government is established (even though clear majorities in free countries support an institutionally and/or functionally specified world government—just less so than citizens of partly/not democratic countries). These results reflect findings with respect to public support for existing IOs, where it has been found that citizens of less developed countries, who perceive greater corruption in their country, trust the UN more—possibly due to the hope for a compensating effect (Torgler 2008).
Material power is consistently negatively associated with world government support across all experimental conditions using equal and population weights in our various regression analyses, which is in line with Hypothesis 9b (see online appendix, tables AT47–AT52). Exceptions are again the COVID-19 conditions, where the associations are generally insignificant and even positive in two conditions. Furthermore, figures AF4–AF7 in the online appendix show that the difference between respondents from more and less powerful countries appears to be more linked to the functional than the institutional specification. This indicates that concerns regarding national sovereignty over policy areas may help explain the divergence between citizens of more and less powerful countries. One reason for the largely negative association may be that citizens of stronger countries tend to fear that the establishment of a world government would lead them to lose their relatively privileged international status. Conversely, citizens of weaker countries may see the establishment of a world government as a potential way of evening out power inequalities in the international system and thereby improving their status in the global order (cf. Scheuerman 2011).
Finally, let us consider results in individual countries. Figure 6 summarizes attitudes in each country toward a democratic world government focusing on global issues broadly, while online appendix 2.3 presents results for the remaining experimental groups in each country. First and foremost, note that majorities in all countries—except for the United States—support the proposal of a fully specified world government. Egypt, India, Kenya, Indonesia, South Korea, Columbia, and Hungary have the largest majorities in favor of a fully specified world government, ranging from 75 to 82 percent of respondents supporting the idea. The diversity of these countries—among others, in terms of our theorized factors of population, development, freedom, and power—illustrates the broad appeal of the world government proposal across the world. The least supportive nations—apart from the United States—are Russia and Argentina, where support for the fully specified proposal is at 56 and 58 percent, respectively—still comfortable absolute majorities.
Attitudes toward a fully specified world government, by country. Notes: Potential deviations from 100 percent in each row are due to rounding. The data underlying this plot comes from all three main survey rounds (see online appendix, table AT1).
Aside from the democratic and global issues specification, all institutionally and/or functionally specified proposals are highly popular internationally, with top levels of national support at over 80 percent for each of them (see online appendix 2.3). Majorities in only three countries (Russia, the United Kingdom, and the United States) oppose one or more of the specified world government proposals (see online appendix, figures AF23, AF27, and AF28). However, the unspecified world government proposal is rejected by majorities in eleven of our seventeen survey countries (see online appendix 2.3).
The most opposed outlier is the United States, where all proposals are rejected by majorities, ranging from 55 percent against the democratic and global issues specification (see figure 6) to 70 percent against an unspecified world government (see online appendix, figure AF28). One possible reason—in line with our findings on power as a potential explanatory variable and further evidenced by relative public skepticism in Russia (see online appendix, figure AF23)—is the United States’ exceptional status as the world’s primary superpower.5 The high proportion of item nonresponses in the United States (see table 1) adds to an impression of American exceptionalism in this respect, meriting further investigation. For now, we can conclude that the claim that a world government is or would be rejected by a popular majority (Hypothesis 1) applies only in one country that we study, showing that extrapolations of this specific national context and the average American’s unusually solid opposition to world government would be unwarranted. Moreover, recent survey-based research has shown that majorities of Americans support ideas such as “global democracy,” including a world parliament and global government (Ghassim 2020), or a “supranational organization” with the power to make binding decisions on global challenges (Global Challenges Foundation 2017; ComRes and Global Challenges Foundation 2018; Novus and Global Challenges Foundation 2020). Thus, whether majorities of Americans endorse the establishment of a global government also appears to depend on the specific wording used to present the idea.
Conclusion
Our findings indicate widespread international public support for the strongly globalist proposal of a world government. This result is particularly surprising at the present time, as media outlets, researchers, and policymakers concentrate on the resurgence of right-wing nationalism, authoritarianism, and populism in societies all over the world (Norris and Inglehart 2019). Our study thus reveals a largely overlooked side of contemporary global public opinion: majoritarian support for much stronger global governance institutions than those that currently exist. These findings are especially relevant at a time when the world faces major transnational challenges such as wars, pandemics, poverty, mass migration, and environmental degradation. The apparent inadequacies of our existing global governance system in addressing these challenges suggest the need for fundamental political change at the global level. As country leaders debate suitable policy responses in the short term, our study shows that institutional solutions for the longer term—in the form of sweeping transformations of global governance—enjoy widespread public support around the world. Therefore, political leaders may feel emboldened to pursue far-reaching reforms of global governance—even to the extent of creating a democratic and functionally focused world government—in their pursuit of addressing global challenges effectively.
Our findings indicate that those who pursue stronger and/or more democratic global governance would receive much popular endorsement. Hence, IOs like the UN that have embarked on reform processes (United Nations 2020), as well as civil society organizations advocating different kinds of global governance transformations (Global Assembly Team 2022; Democracy Without Borders 2023; Orback 2023; WFM/IGP 2023), may feel reassured in their efforts. Our surveys indicate that they have strong popular mandates from various nations around the world to pursue the visions that they have been proposing for decades.
Where would public support be strongest if and when the issue of a global government becomes more salient in world politics? Our analysis indicates that citizens of more populous, less free, less powerful, and/or less developed countries are most supportive of such ideas. However, we also show that public endorsement is by no means limited to such countries, but indeed spreads all over the world. While this is good news for global governance reform advocates, our results also indicate where the greatest public resistance lies, namely, in less populous, freer, more powerful, and/or richer countries. Notably, public opinion in the United States constitutes a potential obstacle to any efforts for the establishment of a world government. Global governance reformers thus face the task of having to convince policymakers in countries worldwide—and especially in the United States (Held 2010, 303)—that the establishment of a democratic world government focusing on global issues would be in the best interest of humankind to address the transnational challenges we face.
Funder Information
We are especially grateful to Brian Schaffner for funding our pilot survey on the Amazon Mechanical Turk platform in the context of our Advanced Survey Data Analysis and Survey Experiments class at the Essex Summer School 2017. Moreover, we want to express our immense gratitude to the Department of Politics and International Relations at the University of Oxford, Mathias Koenig-Archibugi, and Luis Cabrera, as well as Jonas Tallberg and the Legitimacy in Global Governance project (Riksbankens Jubileumsfond grant M15-0048:1), for supporting our main survey rounds in Argentina, Australia, Canada, China, Colombia, Egypt, France, Hungary, India, Indonesia, Kenya, Russia, South Korea, Spain, Turkey, the United Kingdom, and the United States. Lastly, we are indebted to Yale-NUS College and Dublin City University for generous financial support.
Footnotes
Our preregistration document is available here.
The full questionnaires are available as supplementary materials on the ISQ Dataverse.
The exception here is the United Kingdom (UK), as YouGov includes explicit “I don’t know” choices by default. Since we count “I don’t know” as item nonresponses, this helps explain the higher proportion of no-opinion responses in that survey (see table 1).
The replication code and data for our analyses are available on the ISQ Dataverse.
While the NMC scores strongly draw on population variables and thus present China as significantly more powerful than the United States, our characterization of the United States as the world’s primary superpower is in line with common conceptualizations and interpretations (Monteiro 2014).
Author Biography
Farsan Ghassim began this project as a DPhil in International Relations student at the University of Oxford, further worked on it as a Postdoctoral Fellow at Lund University, and completed it as the Junior Research Fellow in Politics at The Queen’s College in Oxford.
Markus Pauli is Assistant Professor in political science at Dublin City University. He has held positions at the National University of Singapore, Yale-NUS, and Heidelberg University. He co-authored “Statecraft & Foreign Policy: India 1947–2023,” Dublin City University Press.
Notes
Authors’ note: We are grateful to Brian Schaffner and the Advanced Survey Data Analysis and Survey Experiments class at the Essex Summer School 2017, especially our fellow group members Takaaki Hirotani and Karuna Sriklung. For further useful feedback and input, we thank the organizers, discussants, and audiences at Yale-NUS College in Singapore in October 2018, the Martlets Society at University College Oxford in November 2018, the University of Zurich in March 2019, the Annual Convention of the International Studies Association in Toronto in March 2019, the Annual International Conference of the Political Studies Association in Nottingham in April 2019, the RECONNECT workshop in Münster in January 2020, and the University of Shanghai in October 2023. For their help in finalizing our draft, we are grateful to Josu Aurrekoetxea, Jules Salomone-Sehr, and Jivanta Schöttli. Finally, we want to thank our research assistants (Christian Sigl, Ian Neidel, and Shawn Thacker), the various translators (Benedek Paskuj, Chandrachur Singh, Demirkan Coker, Grace Baghdadi, Ismail Hmitti, Jenny Lee, Jesse Kimotho, Kevin Misaro, Laura Palacio Londoño, Léo Bureau-Blouin, Leonhard Tannesia, Marwan Jalani, Maxine Setiawan, Miklós Szabó, Oksana Levkovych, Pilar Elizalde, Pinar Aldan, Rachel Zhou, Ricardo Aguilar, Sapheya Elhadi, Seebal Aboudounya, Victory Lee, and Yan Tong), and the many pretesters for their valuable contributions to our research project. Our appendix, questionnaires for all survey countries, and replication files are available on the ISQ Dataverse at https://dataverse.harvard.edu/dataverse/isq. Any questions about the replication of our results may be directed to the corresponding author.
