Abstract
Class is often neglected as a factor influencing foreign policy. While recent research explains the foreign policy positions of states in terms of the preferences of a ruling regime’s key constituencies of support, these accounts have not investigated how inter-state relations are influenced by specific class-based social forces. Influenced by liberal pluralism, they are agnostic about the role of class. By contrast, neo-Gramscian approaches conceptualise foreign policy as resulting from the configuration of class-based social forces, which form a ‘state-society complex’ in conjunction with institutions. The foreign policy stances of states have social foundations. Drawing on an expert survey by the Varieties of Democracy project, fixed-effects and first-difference regression analyses indicate that dependence on support from specific class-based groups is associated with distinct voting positions at the United Nations General Assembly. These findings are consistent with the argument that foreign policy has social foundations: class politics shapes world politics.
Introduction
Profound shifts in a ruling elite’s class support, or the conquest of power by new rulers that draw their support from a different coalition of classes, often results in foreign policy change.
As China completed negotiations to join the World Trade Organisation in the early 2000s, Communist Party General Secretary Jiang Zemin advanced the doctrine of the ‘Three Represents’, which involved broadening the social base of the party to include new groups in the private sector (Shambaugh, 2008: 115). In Chile, Salvador Allende’s presidency accorded official status to unions, supported new worker councils (Zapata, 1976), joined the Non-Aligned Movement and established diplomatic relations with Cuba. These foreign and domestic policies were reversed after Pinochet’s coup in 1973, which brought a military regime supported by business elites and technocrats to power. Increasingly dependent on a communist party drawing its support from the Javanese rural poor, President Sukarno of Indonesia denounced the NEKOLIM (neocolonial imperialist) powers and, in 1965, withdrew the country from the United Nations (UN). After Suharto’s coup and the violent demobilisation of the populace, Indonesia returned to the UN and rebuilt diplomatic bridges with the West 1 . Subsequent to a successful campaign for democratisation led by students, labour and white-collar workers (Yun, 1997), South Korea left the Group of 77 and joined the Organisation for Economic Co-operation and Development in 1996.
Almost all rulers rely on the continued support of certain social groups to remain in power. The leadership of a state must therefore pay attention to the preferences of constituencies in terms of both domestic and foreign policy. Because class-based social groups are of enduring political relevance in the majority of societies across the world, the organised disagreements between states in institutions such as the United Nations General Assembly (UNGA) do not only reflect divisions between democratic and authoritarian states, or the Global North and South, but also reflect class divisions playing out on a global stage. In this sense, world politics is class politics.
This article provides a test of this claim through a systematic investigation of the association between class support and the voting positions of states at the UN General Assembly, from 1946 to 2020. In doing so, it contributes to two research traditions concerned with the subnational influences on international organisations and the political stances that states adopt within them: liberal pluralism and Marxist historical sociology.
Liberal pluralism holds that the foreign policies of states reflect coalitions of diverse social interests that vie for influence within each state’s framework of political institutions. But despite this ‘bottom-up’ account of the sources of foreign policy stances, liberalism offers no developed account of the foreign policy preferences of different classes. Thus, although many advances have been made in the study of foreign policy positions from within a liberal pluralist framework, discussion of class in these accounts is often notable by its absence. This article challenges these omissions.
A concern with class in world politics is, of course, associated with Marxist and structuralist approaches to International Relations, which typically interpret events in world politics in terms of conflicts between property-owning elites and popular classes. Some Marxist accounts, however, are committed to the position that political institutions and the rule of law are mere tools of class rule, as well as to an account of an inevitable and polarising conflict between workers and property owners. However, when examining particular historical episodes, Marxist scholars have been more flexible, acknowledging sources of social power beyond class and the diversity of class alliances. With these considerations in mind, this article takes a position closer to neo-Gramscian analyses, according to which state institutions and coalitions of class-based social forces together form a state-society complex, giving rise to a variety of forms of state that pursue different ambitions within international institutions (Cox, 1981: 126).
Despite a long-standing research programme investigating the politics of multilateral institutions from a neo-Gramscian perspective, the argument that coalitions of class support underpin positions of states within these organisations has not been evaluated systematically. Quantitative researchers analysing patterns of foreign policy disagreement in international institutions have set to one side the arguments advanced by critical scholars, rather than investigate them (Voeten, 2021: 20). As this article will show, however, class-based political economy approaches to foreign policy present a challenge to the dominant liberal pluralist approach.
In the article, engagement with these debates supports a set of hypotheses about how dependence on the support of different class-based groups affects the foreign policy that a regime pursues, operationalized in terms of voting record in the UNGA. Because states vote annually on dozens of resolutions covering a range of issues on which there is disagreement, voting in the UNGA offers a barometer for assessing diplomatic and ideological support for the status quo in the form of US hegemony and the ‘liberal international order’ (Bailey et al., 2017: 431). The hypotheses stipulate that regimes dependent on property-owning elites will have a greater tendency to vote alongside the United States and that those dependent on working-class groups will have a greater tendency to vote against the United States, but that these tendencies are contingent on what coalitions that class-based groups have entered into.
To investigate these hypotheses, the article draws on data from the Varieties of Democracy project (V-Dem) based on a survey of country experts regarding the regime dependence on different social groups (Djuve et al., 2020). Expert surveys have been used in existing contributions to the study of International Relations, Braumoeller for example utilises a survey of diplomatic historians regarding foreign policy preferences of major powers (Braumoeller, 2012: 86). The V-Dem dataset on regimes of the world provides a means to evaluate arguments about the relationship between the class character of regimes and their foreign policy in a systematic fashion.
The hypotheses are tested using a fixed-effects multiple regression model and a first differences model. Using a fixed-effects model allows us to compare countries with themselves across time, revealing that periods in which ruling groups depend on different classes for support are associated with periods in which they vote differently at the UNGA. The use of a first differences model demonstrates that the shifts in class support coincide with shifts in UNGA voting position. Because approaches such as neo-Gramscianism see the influence of class-based social forces as historically contingent rather than invariant, a rolling regression is performed to evaluate changing relationships over time, revealing that class support remains associated with specific voting positions even in recent sessions of the General Assembly.
The analysis demonstrates that class is an unduly neglected category in the study of International Relations, especially in quantitative research on international institutions, where liberal pluralist assumptions have predominated. The findings have broader implications for empirical research focused on multilateral institutions, as they suggest that factors emphasised by critical and materialist traditions of international theory may have been overlooked due to a particular, and contestable, set of assumptions about the relationship between the state and civil society. Explaining the social origins of foreign policy requires a reengagement with the insights of a class-based analysis of politics – both domestic and international.
Theory and empirical research on class and international disagreement
Liberalism: agnostic pluralism
Much of the systematic research on the foreign policy positions of states and disagreement in international organisations has been conducted within a liberal pluralist framework. Voeten’s Ideology and International Institutions explains patterns of disagreement in multilateral institutions in terms of international ideological conflict. Proponents of different ideologies have ‘fundamentally different views on how domestic societies and international society should be organized’ (Voeten, 2021: 23). The study attempts to demonstrate that the politics of multilateralism can be understood in terms of a single dimension of ideological agreement/disagreement with the US-led international order. Liberal constitutionalist states, conservative governments and states committed to open markets tend to be more supportive of this order (Voeten, 2021: 38). The relationship between class and different ideological positions in world politics is not investigated.
Braumoeller offers a nested, systemic model of International Relations, according to which great powers compete for both shares of power and to shape the ideological character of the system. Each state’s foreign policy goals are shaped by the preferences of the key elements of its citizenry, channelled through the institutions of specific political systems (Braumoeller, 2012: 32, 35, 72). While Braumoeller emphasises competition between liberal and illiberal states, left unexplored is the issue of what sort of constituencies possess what sort of ideological preferences. Similar to Braumoeller’s nested model, research within the Change in Source of Leader Support (CHISOLS) project argues that every leader rests on the support of domestic groups with their own foreign policy interests (Mattes et al., 2016). They find that, conditional on the configuration of political institutions, shifts in a leader’s coalition of support are correlated with changes in a state’s voting patterns at the UN (Mattes et al., 2015). The creators of CHISOLS do not investigate the actual content of different social groups’ interests, as ‘there is significant variation in the relevance of particular social cleavages across countries’ (Mattes et al., 2016: 260).
Although these investigations are compelling, they all remain agnostic about the social sources of foreign policy preferences. This agnosticism reflects the influence of liberal pluralist model of politics. According to Dahl (1971), states with free and fair elections can be characterised as ‘polyarchies’, governed by shifting coalitions composed of many social groups and interests. Moravcsik provides the definitive outline of liberal pluralism in International Relations, characterising it as ‘a “bottom-up” view of politics in which the demands of individuals and societal groups are treated as analytically prior to politics’ (Moravcsik, 1997: 517).
The groups composing these coalitions might be class-based groups, but they could equally be groups organised around any other common interests or values. Liberal political economy does, however, offer an account of how distributional conflict arises from the gains and losses incurred from participation in world markets (Moravcsik, 1997: 519). For example, import-competing industries and groups of actors possessing locally scarce factors of production are expected to oppose free trade. Narizny (2003) argues that different classes may have different preferences about rearmament, but in his analysis of the sources of British and American grand strategy, he emphasises the role of sectors over classes, as sectors can be wiped out by international events such as loss of access to foreign markets (Narizny, 2007: 18).
Liberal international theory, therefore, is either agnostic about class or tends towards a sectoral account of the sources of foreign policy. While it can take account of classes as one of many societal interests, it rarely does so. As a result, it offers few substantive suggestions about how the influence of social classes might affect a state’s positions on global issues debated at the UNGA.
Marxism: internationalised class struggle
An alternative conceptualisation of the relationship between state and society is offered by traditions of historical sociology that emphasise the role of class. Liberal pluralist perspectives conceive of society as being composed of a large variety of different interest groups. Under conditions of electoral democracy, ‘minorities rule’ in the sense that policy is determined by shifting coalitions of interest groups (Dahl, 1971: 75). By contrast, the social forces tradition focuses on large social groups that achieve a sufficient level of organisation to mobilise and enter into the political arena. Even within electoral democracies, some social groups have systematically greater access to state power. Consequently, the way that a state regulates society within the limit of its territorial authority and extends its influence beyond its borders reflects the interests of some social forces over others. Access to state power is structural and enduring, it does not represent the temporary capture of a neutral set of government institutions by a temporary alliance of interests. Although scholars have identified various groups that can act as social forces in specific social contexts, such as faith groups led by communal elites (Slater, 2010), a long tradition within historical sociology emphasises classes as the most important social forces.
Marxism is of course the approach that places greatest, even overriding, emphasis on class and the centrality of struggles over the means of production. Within a Marxist framework, a state’s foreign policy is rooted in its class character, and must change if the rise of new classes transforms the state. Gramsci (1971) sets out this position clearly: ‘Do International Relations precede or follow (logically) fundamental social relations? There can be no doubt that they follow’ (p. 176). Rosenberg makes a case for a Marxist historical sociology of International Relations, arguing that the ascension of a new mode of production and its corresponding class relations transforms the social character of states and thus the nature of state interests (Rosenberg, 1994). Other scholars build on early 20th century analyses of imperialism, which identified the increasing concentration of capital, the establishment of monopolies and cartels, the dominance of finance and the emergence of a stratum of share-owning rentiers within the bourgeoisie as the causes that underpinned European expansionism (Brewer 1990; Lenin 1970(1916)).
Despite differences between strands of Marxist International Relations, a point of agreement is that foreign policy is shaped by a global conflict between capital and labour in which class-based interests are pursued both nationally and internationally (Selwyn, 2015: 523). Although Marxist International Relations does not necessarily reduce world politics to class politics, the tradition expects struggles between social forces over ownership and control of the means of production to influence the relations between states.
The endurance and international relevance of class politics
Although Marx anticipated many important developments within capitalism, his argument that the distributive demands of workers could not be accommodated within the framework of ‘bourgeois’ representative democracy underestimated the possibility for class compromise (Marx, 2020(1975). But although successful worker revolutions did not transpire within advanced industrialised societies, class cleavages were strong enough in high-income democracies for political scientists to treat it as a truism that ‘the party struggle is a conflict among classes’ (Lipset, 1960: 223). Organised labour and parties of the working class pursued greater redistribution and state intervention, while the numerical size of the working class ensured that even centre-right parties remained committed to the goal of full employment. Economic restructuring due to the crises of the 1970s and the political turn towards neoliberalism, however, transformed the political sociology of industrialised societies. This gave rise to a new accepted wisdom that dealignment was taking place, with class becoming less relevant due to individualization, the decline of organised labour, converging of living standards and social value shifts (Inglehart, 1997: 254–255).
Research has demonstrated, however, that this is a misleading picture and that class divides have endured in high income democracies. Focusing on Britain, Evans and Tilley (2017) demonstrate that material disparities between classes persist and that individuals still retain class identities. They find that occupational class remains the strongest predictor of ideological preferences about ownership and control within the economy, with substantial and very enduring differences in attitudes about redistribution and the role of the private sector between workers and the ‘old middle class’ of managers and business owners (Evans and Tilley 2017: Ch. 4). Similar differences in attitudes between classes regarding inequality-related issues have been found in comparative studies of multiple European countries (Kalmijn and Kraaykamp, 2007; Rennwald, 2020). Class-based differences in voting patterns persist within high-income democracies (Evans and De Graaf, 2013).
Class also remains relevant to voting behaviour in the democracies of the Global South. Class might have become more rather than less important as a factor influencing voting in Latin America in recent decades (Mainwaring et al., 2015). In Indonesia, class measured in terms of income has been robustly related to voting behaviour since democratisation, alongside the country’s important urban-rural divide (Mujani et al., 2018: Ch. 4). Scholars have found that class plays a role in African elections alongside ethnic identity and the urban–rural divide, and have investigated the distinctive preferences of Africa’s rising middle classes (Cheeseman, 2015).
Differences in political preferences arising from class differences, therefore, remain evident across many societies. Cross-national evidence is consistent with the argument that working classes have a tendency to support redistributive policies, the alleviation of inequality and state intervention to restrict the domain of the market, with propertied classes tending to oppose these goals. If foreign policy is the external pursuit of the interests of domestic social forces, then it is plausible to expect that a state’s evaluation of the global political economy will depend on whether propertied classes or workers are more powerful as domestic social interests.
Studies of voting at the UN General Assembly interpret the voting positions of states as reflecting evaluations of US global leadership (Bailey et al., 2017). Since WW2, the United States has implemented a particular vision of international order. An element of this vision has been a commitment to maintaining an open-world economy and an opposition to barriers to trade and investment (Kolko, 1988; Layne, 2007). The overarching ideology of foreign policy driving American administrations has been described as a form ‘liberal absolutism’ thoroughly committed to individual liberties and property rights, with its roots in America’s absence of a feudal past (Desch 2007; Hartz, 1991). Similarly, critical political economists have seen the United States as a thoroughly ‘Lockean’ society, based from the outset on the institution of private property (Halperin, 2004: 33).
The US commitment to markets, property rights and the ‘open door’ has frequently brought it into collision with regimes supported by popular social forces. Within democracies, parties linked to the organised working class have sought nationalisation, the expansion of state intervention and progressive taxation. Revolutionary states established in the name of workers and peasants pursued more radical policies of state control, expropriation, and the repudiation of international debts. Statist policies were also pursued by anti-colonial populists such as Nasser in Egypt and non-socialist, labourist leaders such as Peron in Argentina. Patterns of foreign ownership and investment internationalised some of the resulting distributive conflicts, which linked them to mid-20th century anti-colonial struggles, as post-colonial regimes sought to free themselves of what they perceived as neo-colonialism and dependency (Silver and Slater, 1999). Solidarity was expressed for these goals, and related causes such as the anti-apartheid struggle, by labour movements and socialist parties in the democratic states of the Global North (Gurney, 2009). Support from USSR, and other state-socialist regimes claiming to rule on behalf of the working class, connected these struggles to the East-West Cold War. One result was the ‘widespread use of U.S. military and police power, often operating through clients and proxies, to suppress or eliminate popular forces identified as communist’ as well as overthrow disagreeable governments in the Global South (Barkawi, 2001: 107).
Scholars in the critical political economy tradition have argued that the bourgeoisie in both the North and the South has been supportive of US hegemony. The representatives of landowners and business elites might have already been predisposed to be favourable to the pro-market foreign policies of a capitalist United States, but the threats of redistribution and revolution deepened their sense of insecurity and pushed them further towards Western anti-communist patrons during the 20th century (Westad, 2007). Galtung’s (1971) structural theory of imperialism emphasises the role of ‘comprador’ classes in the periphery, the local ruling classes whose interests are closely aligned with the dominant states of world system. Roos (2019) has recently made a similar argument that local business elites and their representatives play a central role in the enforcement of international debt repayment in the Global South.
In the post-Bretton Woods era, the World Bank and the International Monetary Fund have played a key role as ‘globalizers’, encouraging countries experiencing economic distress to integrate more fully into the global economy (Woods, 2006). The prescriptions and conditionalities of the international financial institutions have often involved reducing state intervention and welfare. Given the consistent preferences among working classes in different societies, states dependent on the support of the working class are likely to be more opposed to these pro-market and anti-redistributive policies, which may be reflected in their evaluations of the US-led liberal international order.
Class alliances, state-society complexes and multilateral institutions
A further challenge for Marxist approaches is that, even if significant differences in preferences between classes endure, advanced capitalist societies have not polarised into a propertyless mass of proletarians and a numerically small bourgeoisie as anticipated. The urban working class has almost never comprised a majority, peaking in North-Western Europe before the middle of the 20th Century (Przeworski, 1986: 39). Since the 1970s, the industrial working class has shrunk in high-income capitalist societies, and the ‘new middle class’ of white-collar service-sector workers has grown (Erikson and Goldthorpe, 1992). Social democratic parties responded to these developments by courting the support of middle-class voters, marginalising the influence of working classes and reducing their substantive representation in high-income multiparty democracies (Evans and Tilley, 2017: 199–200). These patterns are of course not uniform across the world. Through trade and off-shoring, restructuring in high-income economies has been related to the emergence of a new global division of labour and the growth of the urban working class in export-oriented industrialisers such as China (Hallward-Driemeier and Nayyar, 2018: 46–48). In addition, class divisions might be sharpening again within industrialised societies, due to a tendency towards ‘hourglass’ labour markets due to the elimination of many middle-income jobs (Kalleberg, 2013). But despite the enduring global relevance of class divisions, the class make-up of capitalist societies rarely if ever simplifies into the confrontation between a small propertied class and a supermajority of urban workers.
Given the complex, differentiated character of class structures in all societies, the traditional Marxist argument that the state is a vehicle for a single class to exercise domination over others becomes harder to sustain. To rule, class-based social forces will typically need to enter into cross-class coalitions. Marx acknowledges this in his account of the French Second Republic, although he interprets the period of fluid coalitions as a prelude to a final battle between an isolated working class and the bourgeoisie (Marx, 2002(1852)). Acknowledging that the Russian working class lacked the strength to secure revolutionary goals on its own, Lenin advocated hegemony or leadership of the proletariat over the first the bourgeoisie and, subsequently, the peasantry (Boothman, 2012) 2 . This later became the basis of the Bolshevik strategy of revolution in the colonial world, where the urban working class was small.
The travails of Marxist revolutionary movements confirm that classes rarely seize power alone. As the social forces tradition within comparative politics contends, the emergence of varying alliances of class-based social forces helps to explain major upheavals such as revolutions and democratic transitions (Luebert, 1991). Moore (1966) explained the emergence of communist, fascist and liberal states as the result of revolutionary transformations led by worker–peasant, landowner–industrialist, and bourgeois–aristocratic coalitions, respectively. The importance of peasants and small farmers has been noted by scholars who argue that variation in the success of worker’s movements in Europe can be explained in terms of whether they successfully built alliances with social groups in the rural sector (Ertman, 1998). Although the strategy was later abandoned, Chinese revolutionaries such as Lin Biao (1965) saw the peasantry as playing a crucial role in a global struggle to encircle capitalist industrialised nations. The Viet Minh was a broad worker–peasant front, and the Communist Party of Vietnam continued its attempts to maintain peasant support after unification and through the reform period (Thayer, 2009: 3–4). Rural social forces continue to be relevant alongside urban classes, and class alliances are more common than situations in which a single class has sway over the state.
If there are many alternative coalitions of different combinations of class-based social forces, then a variety of different forms of state may exist in world politics, rather than just bourgeois, proletarian and a few residual feudal or agrarian states. This is captured in the concept of a state-society complex, introduced by neo-Gramscian International Relations scholars emphasising the ‘plurality of forms of state’, but also employed within neo-Weberian historical sociology (Cox, 1981: 126; Hobson and Seabrooke, 2001). Unlike within a liberal pluralist framework, the state is not conceived of as a set of institutions separate from civil society, which pre-political social interests compete to control (Moravcsik, 1997: 517). Unlike in some Marxist interpretations, the state is not a machine for a single class to oppress others. Instead, the state is conceived as linked to, and drawing its power from, class-based social forces (Jones, 2011: 25). The state rests ‘upon certain configurations of social-class forces as a result of internal political struggles’ (Hobson and Seabrooke, 2001: 25) and asserts the right to regulate conflicts between different social groups (Rupert, 1990: 339–440). Different social groups have differential access to state power through recruitment into political office and the state bureaucracy, and via what Tilly (2000) calls ‘binding consultation’ (p. 5), which can take various forms such as corporatism or electoral representation. The influence of social groups is conceptualised as long-term and institutionalised, rather than the product of fleeting coalitions as within liberal pluralism.
States vary in terms of the degree that they are ‘embedded’ and porous to the influence of certain social forces, or ‘autonomous’ in the sense of being insulated from social influences and able to intervene within society (Evans, 1995). Many domestically and internationally effective states seem to possess a high degree of bureaucratic capacity and share power with civil society actors (Hobson and Seabrooke, 2001: 259–260). In uncommon situations where class-based factions are at stalemate, a ruler or clique can make foreign policy somewhat autonomously. This scenario is explored in Marx’s account of Louis Bonaparte’s rise, and in Paul’s foreign policy analysis of the ‘neoconservative moment’ and the US invasion of Iraq (Marx, 2002(1852): Ch. 7; Paul, 2007).
But although institutions and ideas are treated as semi-autonomous causal factors within a neo-Gramscian framework, a central place is still accorded to class. Cox offers an outline of the typical process of change within foreign policy and International Relations: ‘Changes in the organisation of production generate new social forces which, in turn, bring about changes in the structure of states; and the generalisation of changes in the structure of states alters the problematic of world order’ (Cox, 1981: 126). An established strand of neo-Gramscian research analyses the politics of multilateral institutions in terms of attempts to extend the successful intellectual and political leadership of a class beyond national borders (Gill, 1993). In other words, the politics of international organisations involves the assertion and contestation of hegemony. This application of Gramsci’s ideas to International Relations holds that hegemony involves a struggle to establish a ‘collective political subject’ within civil society, a contest over the state and ‘in a final moment, the clash of hegemonically constructed states in competition on the international terrain, in a geopolitical repetition of the originary domestic process’ (Thomas, 2013: 22).
From within this framework, scholars have argued that the United Sates has used multilateral organisations to extend capitalist markets across the globe, while maintaining a minimal but sufficient degree of consent from other states, institutions such as the General Assembly providing a forum for the airing of grievances against US hegemony, in a manner that does not directly threaten the basis of American material power (Murphy and Augelli, 1993; Murphy, 1994; Puchala, 2005). Given the cross-national evidence about class differences in support for state intervention versus the expansion of the domain of capitalist markets, we might therefore expect that, within these institutions, states drawing on the support of different coalitions of classes will take different stances on the desirability of US hegemony. Although their theory of the state is different, neo-Gramscian scholars agree with Voeten in Ideology and International Institutions that what occurs in organisations such as the UN General Assembly is ideological contestation, with states articulating and attempting to build support for their preferred visions for world order, their positions shaped by the domestic constituencies of support they are beholden to. The difference is that, unlike liberal pluralist perspectives, the neo-Gramscian perspective does not regard the influence of classes on this process of ideological contestation as marginal, temporary, or idiosyncratic, but central and enduring.
Hypotheses on class, foreign policy and international disagreement
The implications of a class-based perspective on multilateral politics can be tested using data from the UNGA. Alongside debates over human rights and specific international crises, many General Assembly resolutions relate to the regulation of the world economy or the issue of development. At certain points, economic issues have been some of the most high-profile, such as during the debates in the 1970s over the New International Economic Order and the Charter of Economic Rights and Duties of States, which were attempts to reorganise the global economy to benefit the states of the Global South and to regulate the activities of multinational corporations, respectively. Moreover, economic and redistributive issues have often been entangled with North–South and East–West controversies in world politics, as discussed above. As a result, we might expect to find a pattern of disagreement between regimes depending on the support of propertied classes and regimes representing ‘popular forces’ within the UN.
This is the central thesis of the investigation, but qualifications and alternative hypotheses must be considered. As outlined above, research in the social forces tradition has repeatedly concluded that the distinctiveness of the peasantry and rural sector must be taken into account for class-based explanations to be tenable. States dependent on the support of rural as opposed to urban workers might adopt different foreign policy positions, owing to the distinct interests and worldviews of the rural poor. Amin has argued that the present world economy is fundamentally antithetical to the livelihoods of billions of peasants, who are being driven to pauper status by a monopolistic global agribusiness industry (Amin, 2012). Substantial evidence exists that the values that predominate in a rural social setting are very different from those of industrialised, urban societies (Heinrich et al., 2010).
Within comparative politics, the middle classes are sometimes regarded as having a moderating role and a preference for the status quo. In Political Man, Lipset argues that ‘the party struggle is a conflict among classes’, but a large middle-class tempers this conflict by inducing political parties towards moderation (Lipset, 1960: 50, 66, 128–30). If the middle class is inclined to support the status quo as well as the institutions of both property ownership and representative government, then its representatives might pursue foreign policies that are supportive of the existing international order and the leadership of a liberal, capitalist superpower.
Rueschemeyer et al. (1992) argue that in many historical cases, the achievement of liberal democracy has represented a class compromise in which the organised working class and its political representatives moderated their redistributive demands, for fear of overthrow of democracy by property-owning classes 3 . The decision to compromise with property-owning classes out of political necessity might blunt the radicalism of working-class political movements; from a neo-Gramscian perspective, this could be interpreted as co-optation and the exercise of hegemony by capitalist classes. States that depend on such a cross-class coalition may be less likely to pursue radical foreign policies that challenge the international status quo, less likely to resist a market-based global order.
The traditions of research outlined above agree that states and governments are responsive to the demands of powerful groups within civil society, anticipating a connection between support from class-based groups and a state’s foreign policy. Marxist and structuralist accounts of International Relations expect that social classes favouring the protection of private property will tend to have a preference for a foreign policy aligned with that of the pro-capitalist United States. Social classes favouring redistribution will tend to have a preference for a foreign policy that would put a regime at odds with the United States within the UN.
H1: The greater the extent to which a regime depends on the support of economically powerful property-owning classes, the more likely it is to vote alongside the United States and its close allies in the UN General Assembly.
H2: The greater the extent to which a regime depends on the support of the middle classes, the more likely it is to vote alongside the United States and its allies in the UN General Assembly.
H3: The greater the extent to which a regime depends on the support of the rural or urban working class, the less likely it is to vote alongside the United States and its allies in the UN General Assembly.
These perspectives anticipate that drawing on the support of different will vote in different ways at the UN, with states supported by property-owning classes voting alongside the global capitalist hegemon in support of the status quo. But the Gramscian tradition, as well as other perspectives in comparative politics and historical sociology, emphasise the importance of the coalitions that classes enter into. Therefore the following hypotheses should also be investigated:
H4: If a regime depends on the support of a coalition of workers and business elites, it is more likely to vote alongside the United States and its allies in the UN General Assembly.
H5: If regime depends on the support of a coalition of workers and peasants, it is less likely to vote alongside the United States and its allies in the UN General Assembly.
Data and methodology
The dependent variable for the analysis is Bailey, Strezhnev and Voeten’s estimate of the ideal point of states in terms of the primary dimension of disagreement at the UN General Assembly (Bailey et al., 2017). The voting record at the General Assembly provides a useful way of ascertaining the ideological positions of states with regards to controversial issues within international society, both during the Cold War and after (Voeten, 2004). This is because many prescriptive resolutions, covering a wide range of issues relating to controversies in the global order, are subject to contested votes in each General Assembly session. The process of setting the agenda by drafting and sponsoring resolutions is not controlled by a few major powers, but is shaped by large coalitions of like-minded states such as the European Union and the Group of 77 (Mesquita et al., 2022: 1244–1245). States are sensitive to the wording of resolutions they vote upon, as resolutions can potentially constitute soft law (Roberts and Sivakumaran, 2018: 102–103), and therefore play a role in establishing and modifying international norms. Voeten interprets this continuous ideal point measure as indicating how far a state supports, or is opposed to, American global leadership and the institutionalisation of liberal and market-based principles within the international order (Voeten, 2021: 23–24).
The focal independent variables measuring regime dependence on the support of a particular class are drawn from the investigation by Djuve et al. (2020) into patterns of regime breakdown as part of the Varieties of Democracy Project (Lindberg et al., 2014). Each state with a functioning government is coded as being governed by a regime that depends on the support of certain key social groups or constituencies. For each social group, a variable indicates whether the group is supportive of the incumbent regime within a country and whether their withdrawal of support would be consequential for that regime’s survival. Support for a regime relates to support for a political system rather than any particular government. As a result, the values are largely stable apart from periods of political transition.
For this investigation, variables measuring dependence on support from the aristocracy, agrarian elites, business elites, the rural middle class, the rural working class, the urban middle class and the urban working class were used. There is a close fit between the class categories used within the VDem project and the theoretical framework developed above. Social classes are distinguished in terms of the extent of their property ownership and in terms whether they are an urban or rural group. Groups are not just categorised as being propertied or propertyless, middle classes are acknowledged as potentially distinct intermediary groups 4 .
Admittedly, these categories are coarse and do not, for example, differentiate between members of the middle-class performing professional and managerial work or among fractions of capital. Nonetheless, as the project records the level of dependence of regimes on seven different class-based groups, the data provide more comprehensive and detailed evidence about the class-based character of UN member-states than has previously been available.
The values are based on the judgement of the country-specific experts consulted, aggregated using Bayesian item-response techniques. The characterization of the class-character of different regimes is not uncontroversial; during the Cold War a debate raged over the question of whether the Soviet Union was an authentic proletarian state, a deformed worker’s state, state-capitalist or a totalitarian state governed by a ‘new class’ (Djilas, 1957). Nonetheless, in terms of face validity, the measures reflect relevant differences in the key constituencies of support of different states. Established liberal democracies tend to be coded as relying on the support of several different classes. Autocratic states tend to be coded as relying less on class-based groups, and on fewer groups. Such states can be thought of as ruling despotically over society, relying on their security apparatus rather than drawing their power from civil society and representing broad class-based social forces. Nevertheless, one-party state socialist polities tend to be coded as relying to some extent on the support of working classes. In this specific respect, communist regimes such Vietnam are coded as sharing similarities with democracies such as Denmark and New Zealand. We should expect to find cases such as these if institutional and class support variables measure distinct attributes of polities, and if class and institutions are not reducible to one another. The Kingdom of Saudi Arabia, by contrast, is coded as dependent on the support of the aristocracy, as is Nepal until the 2008 abolition of the monarchy.
As this last example suggests, the dataset seems to track major shifts experienced by individual states. It records a major shift in the basis of class support of the state after the Allende government in Chile was overthrown and the Pinochet dictatorship established in 1973, with the new regime no longer depending on the support of the working class at all and relying instead on business elites. The civilian regime that succeeded the military regime in Brazil in 1985 is coded as switching from dependence on agrarian elites to dependence on urban classes, while continuing to rely on business.
Figure 1 illustrates patterns of change in the class support since 1946. Most notable is the large increase in the proportion of regimes depending on support from business elites during the 1980s and 1990s. Dependence on support of the middle class also grew on average in the same period but seems to have fallen somewhat since 2000. Nonetheless, from 1990 business elites and the urban middle class stand out as the most important support groups since. The data indicate that in the late 20th century the ‘forward march of labour’ lost momentum (Hobsbawm, 1978), as the proportion of regimes relying on the support of the urban and rural working classes first grew in the post-war period but then plateaued. The worldwide relevance of aristocracies and agrarian elites fell almost continuously. The post-Cold War period could therefore be said to be a distinctively bourgeois era.
Time series line graph of average dependence on the support of different classes 1946 to 2021.
Any apparent statistical correlation between class-based support and voting position at the UN could potentially be epiphenomenal, a spurious association that arises because of factors that influence both. For this reason, measures of factors that could plausibly influence both the political influence of classes within a state and that state’s foreign policy should be included in our analysis.
Findings of previous studies indicate that liberal institutions of representative government are related to a more pro-US voting stance by a state (Kim and Russett, 1996; Voeten, 2021). This argument needs to be addressed in our investigation, because existing traditions of research argue that the configuration of political institutions is itself related to the political influence of classes. Lenin claimed that parliamentary democracy is the ‘best political shell’ for bourgeois rule (Lenin, 1933(1917) Ch.1S.3), and Moore (1966) insisted ‘no bourgeois, no democracy’ (p. 418). By contrast, Rueschemeyer et al. (1992) argue that the working class is the most ‘consistently pro-democratic force’ across different national contexts (p. 8). What unites these perspectives is agreement that political institutions are related to the preferences and strategies of organised social classes, although the notion that there is perfect congruence between class influence and a specific set of institutions is an unpersuasively mono-causal explanation. Furthermore, the relative power of different classes may influence the political system of a state, but electoralism often pushes elites to court popular classes (Collier, 2000). This means that the institutions of representative government might be related to both class influence and a state’s voting stance at the UN, constituting a potential confounder. The Varieties of Democracy ‘polyarchy’ measure of free and fair elections is therefore included in the regression models as a control variable, as is the ‘liberal’ index measuring the presence or absence of liberal constitutionalism and restraints on the executive 5 .
Economic development is closely related to the formation of classes, but might also serve as a proxy for a state’s satisfaction with its position in the world-system or its preferences with regards to global distributive issues. It is therefore included as a control variable. Investigating the ‘capitalist peace’ argument, Mousseau (2003) finds that democracies only tend to vote alongside one another at the UN if they are of similar income levels. For this reason, the interaction between polyarchy and economic development is included in the regression models.
The analysis utilises fixed-effects ordinary least squares regression as the primary estimation strategy to investigate the association between class-based variables and ideal point within the UN General Assembly. Country-specific effects are included in the models to take account of unobserved country-specific confounders, such as enduring institutional or cultural attributes not captured by the independent variables. Year-specific effects are included to take account of shocks and trends across the international system. Adopting a fixed-effects approach affects the interpretation of the results: the regression model evaluates whether correlation exists between values of the independent variables above (or below) the average for a country and values of the dependent variable above (or below) the average for that country, adjusting for tendencies in a specific year. This allows us to assess whether individual countries vote differently in different periods depending on the coalition of class support that the regime in that period depends on. Additional insights are provided by a first differences regression, regressing changes in UN ideal point on changes in the independent variables. This allows us to evaluate whether changes in the class basis of regime support coincide with changes in a state’s foreign policy position.
Analysis
Table 1 reports bivariate correlations between the dependent variable, UN General Assembly ideal point closer to the United States, and the independent variables. As the table shows, support for a position aligned with the United States at the UN General Assembly is correlated positively with polyarchy, liberal constitutionalism, economic development and dependence on all classes apart from the aristocracy. The class support variables are generally correlated positively with one another, although the association between support from agrarian elites and support from urban and working classes is weak and support from the aristocracy correlates negatively with dependence on other classes. Of course, bivariate correlations do not take account of how the independent variables might confound one another.
Matrix of correlations.
Moving to fixed-effects ordinary least squares, Model 1 in Table 2 shows that dependence on support from business elites and the rural middle class is associated with a more pro-US voting position, while dependence on support from the rural and urban working classes is associated with the opposite. This corroborates H1 and H2 rather than H3 and H4. Even though it is descriptively true that regimes depending on the rural and urban working classes have a greater tendency to vote alongside the United States, once the effect of political institutions, economic development and dependence on other classes are taken into account, support from these groups is negatively associated with agreement with the United States. H1 is only partially corroborated, however, as there is no significant association between support from the aristocracy or agrarian elites and voting position. The null hypotheses cannot be rejected with regards to H4, as despite the strongest bivariate correlation with the dependent variable among the class support variables, there is no significant association between urban middle-class support and voting position at the UN when controlling for other factors. In terms of the control variables, polyarchy is only positively associated with agreement with the United States when levels of economic development are high, in line with the capitalist peace argument. Liberal constitutionalism, however, has a positive association with agreement with the United States 6 . Different aspects of liberal democracy—elections and the rule of law—seem therefore to be associated with different voting tendencies at the UN.
UNGA ideal point and class-based support, 1946–2020.
Cluster-robust standard errors in parentheses.
p < 0.01, **p < 0.05, *p < 0.1.
To aid interpretation, Figure 2 displays estimated effects of the class support variables on voting position in the UNGA, based on Model 1. For comparison, the coefficient plot also includes estimated effects when no other control variables apart from country and year-specific dummy variables are included. Model 2, using first differences regression, shows that changes in class support are robustly associated with changes in voting position in the UNGA. Results are very similar to the fixed-effects model, although changes in whether a regime depends on support from agrarian elites is significant in the first differences model.
Coefficient plot for estimated bivariate and multivariate effects on dependence on the support of different class-based social groups on voting position in the General Assembly. Country-specific and year-specific effects included in all models.
A question might be raised whether these methods are appropriate, given that the theoretical traditions this article draws upon challenge the notion of invariant social relationships. Scholars of comparative political sociology have expressed scepticism of the idea ‘that there is a homogeneous pattern of causation throughout history’ (Rueschemeyer et al., 1992: 284). Similarly, neo-Gramscian scholars criticise ahistorical approaches that fail to take account of qualitative changes in the character of states and world order, assuming ‘a continuing present’ (Cox, 1981: 129). Despite research on political attitudes indicating that class is of enduring relevance, sceptics might still question the importance of class-based social forces in post-industrial societies, or doubt the influence of expanding urban working classes in contemporary emerging economies. Class politics is no longer as closely linked to superpower tensions as it was during the Cold War, and the worldwide embrace of market-based development strategies might have brought about a form of international dealignment, whereby international economic issues that propertied and popular classes previously had strong preferences about are no longer subject to as much international contestation within the General Assembly. To address these concerns, a panel rolling regression was estimated with a window of 5 years either side of each session of the UN General Assembly 7 . Rolling regression re-estimates the same regression model for successive time periods, revealing changes in model coefficients over time. The advantage compared to the fixed effects approach is that rolling regression reveals how the relationship between class and voting has changed over time, the disadvantage is that it precludes within-country comparisons over long time periods.
Figures 3 to 6 display the changing associations between UN ideal point and business, rural middle-class, rural working-class, and urban middle-class support. The solid lines represent estimates of the effect of a greater value for the relevant class support variable than within-region averages, within a specific time window around each year. Dashed lines are 95% confidence intervals. The visualisations show that all four class support variables had large and significant associations in the 1960s and 1970s. Differences in class support seems to have become less-strongly associated with foreign policy positions during the final years and aftermath of the Cold War, although support from business and from the urban working class remained statistically significant and has even rebounded in substantive significance in recent years.
Changing estimated effects of dependence on business elites, based on a multivariate rolling regression, with 95% confidence intervals.
Changing estimated effects of dependence on the rural middle class, based on a multivariate rolling regression, with 95% confidence intervals.
Changing estimated effects of dependence on the rural working class, based on a multivariate rolling regression, with 95% confidence intervals.
Changing estimated effects of dependence on the urban working class, based on a multivariate rolling regression, with 95% confidence intervals.
The hypotheses relating to class coalitions and foreign policy can be investigated by interacting specific variables relating to class support, examining interactions separately to avoid multicollinearity. In Model 3, the interaction of dependence on business elites and the urban working class is positively signed, consistent with H4. Regimes receiving support from this multi-class coalition therefore seem to be in greater agreement with liberal internationalism. As Figure 7 depicts, support from the urban working class without simultaneous support from business elites is associated with disagreement with the United States at the UNGA. However, in Model 4, the interaction of rural and urban working-class support is insignificant. A worker-peasant alliance is not associated with especially strong opposition to the US-led international order, although dependence on support from either class is associated with opposition.
Marginal effects plot showing the effect of high dependence on support from the urban working class, for different levels of dependence on business elites.
Overall, the results are broadly consistent with the argument that the conflict between capital and labour, between propertied and working classes, plays itself out in International Relations. Support from the aristocracy does not seem to contribute to an explanation of foreign policy measured in terms of UNGA voting positions, however, and evidence that support from agrarian elites is correlated with a more pro-US voting position was mixed. Periods in which regimes are more dependent on the support of the urban middle class were not associated with a more pro-US position at the UNGA in any specification. Despite the importance of urban middle-class support for many states, it is support from other classes that is associated with distinctive voting tendencies at the UN.
The results were consistent with the argument that regimes depending on working class support are contingent supporters or opponents of the US-led international order. Regimes depending on working-class support tended to vote in opposition to the liberal international order if their coalition of support does not also depend on business elites. The foreign policy position of a regime therefore seems to depend, not just on the class-based social forces they draw their support from, but the way in which classes are incorporated into the political system and the coalitions that they are part of.
Conclusion: class counts in world politics
This article has aimed to advance the empirical study of the role of class in world politics and the investigation into the domestic underpinnings of a state’s foreign policy. Existing research has established that states governed by different political institutions vote differently in the UNGA and that shifts in the coalition of support of a state’s leader are associated with changes in voting patterns. But this strand of research has not examined the relationship between support from class-based social forces and voting position in the UNGA, nor has it explored the social underpinnings of foreign policy ideology in the preferences of social classes, channelled through institutions.
To address this, the investigation within this article has attempted to advance systematic, quantitative research on the relationship between domestic class politics and the foreign policies states pursue within a multilateral context. The findings confirm aspects of both liberal and Marxist International Relations theory. The ‘bottom up’ approach of liberalism is corroborated by the results, but the argument that little can be said about the social groups that support particular foreign policy stances, because their preferences vary idiosyncratically from country to country, has been found to be needlessly agnostic. Support from particular classes has been found to be robustly associated with particular voting stances at the UN. Ignoring class represents a major paradigm-driven oversight by liberal pluralist International Relations theory.
The Marxist or structuralist view that foreign policy has a class character is corroborated by the findings. Sidelining the arguments of these traditions in the analysis of foreign policy and the politics of international institutions is unjustified. Nonetheless, while support from business elites is consistently associated with a US-aligned position in the UNGA, the statistical connection between support from the urban working class and foreign policy alignment is a contingent one. As implied by neo-Gramscianism and sociological approaches to comparative politics, different class coalitions may result in state-society complexes with very different characters and foreign policy tendencies.
Several avenues for further research suggest themselves. The relationship between class support and other aspects of foreign policy could be investigated, including arms transfers, alliance portfolios and international organisation membership. As noted, liberal theory identifies sectors as key interest groups influencing a state’s foreign policy. Data on sectoral support for regimes would enable their influence to be contrasted and compared with classes. In addition, this article has considered the influence of class relations within states on inter-state relations, yet the transnational dimensions of class were emphasised by neo-Gramscian scholars from the outset of the research programme (Cox, 1981: 147). Nonetheless, the results advance this research agenda by demonstrating robust associations between class politics within societies and the politics of multilateral institutions.
The analysis in this article challenges the liberal pluralist tendency to remain agnostic about the influence of class-based social forces on foreign policy, and to sideline arguments from critical political economy and historical sociology. While it is important to acknowledge that voting at the UNGA is not simply the diplomatic translation of the class struggle, the indirect influence of classes with different interests and ideological evaluations of the contemporary world order has been overlooked in recent research on international organisations. The character of International Relations at a given point in time may depend on the classes that are incorporated into different state-society complexes within the international system. World politics may be more than class politics alone, but class politics do seem to shape world politics, as neo-Gramscian scholars have long maintained.
Supplemental Material
sj-txt-1-cac-10.1177_00108367241269623 – Supplemental material for Is world politics class politics? States, social forces and voting in the United Nations General Assembly 1946–2020
Supplemental material, sj-txt-1-cac-10.1177_00108367241269623 for Is world politics class politics? States, social forces and voting in the United Nations General Assembly 1946–2020 by Nicholas Lees in Cooperation and Conflict
Footnotes
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
Supplemental material
Supplemental material for this article is available online.
Notes
Author biography
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
