Data

Our interest here is to illustrate citizen attitudes toward IOs across countries and over time. However, comparative data on public opinion vis-à-vis IOs are in short supply. The only cross-national dataset on citizen opinion toward IOs with global reach is the WVS. The WVS has in recent waves included a growing number of organizations in a question about confidence in IOs, but there are important limitations. With the exception of the UN, the included IOs are almost exclusively regional organizations. Moreover, the countries covered tend to vary from one survey wave to another, making systematic tracking of legitimacy beliefs over time difficult (Dellmuth Reference Dellmuth, Tallberg, Bäckstrand and Scholte2018). Next to the WVS, public opinion toward IOs is covered in a number of regional barometers, but then with an exclusive focus on specific regional organizations. In our mapping of citizen legitimacy beliefs toward IOs, we therefore use a combination of data from the WVS and regional barometers.

First, we use data on the UN from the EVS and WVS in order to map attitudes toward this organization in specific countries over the time period 1994–2020. The UN is the global IO for which we have the most encompassing public opinion data (Online Appendix B).

Second, we complement the analysis of the UN with an inquiry into regional organizations in three geographical contexts with adequate data availability: Africa, Europe, and Latin America. For Africa, we use data from the Afrobarometer on citizen support for the AU over the time period 2002–2015. For Europe, we use data on trust in the EU from the Eurobarometer over the period 2003–2017. In Latin America, we use data from the Latinobarometer on support for Mercosur over the time period 2001–2015.

Third, we examine cross-IO variation in citizens’ legitimacy beliefs based on new data from the WVS7. Thanks to a question battery we contributed to this wave, we have access to comparable data on citizens’ confidence in a broader range of global IOs for the period 2017–2020 (for in-depth analyses using these data, see Dellmuth et al. Reference Dellmuth, Scholte, Tallberg and Verhaegen2022a, Reference Dellmuth, Scholte, Tallberg and Verhaegen2022b). These IOs enjoy substantial authority in their respective issue areas, are reasonably known to citizens, and allow us to assess whether patterns of legitimacy beliefs vary among IOs active in economic governance (IMF, World Bank, WTO) compared to IOs involved in human security governance (ICC, NATO, UN, WHO).

When mapping public opinion toward IOs in this section, we are pragmatic as to the measures used in different surveys. While Chapter 3 explains why we prefer citizens’ confidence in an IO as the measure of legitimacy beliefs, we accept greater diversity in measures in this section, for purposes of being able to identify some patterns based on the scattered data that exist.

Patterns in Legitimacy Beliefs

The UN is a hub and major incubator of ideas in global governance. The organization was established in the tradition of the League of Nations after World War II and has undergone numerous reforms, which have broadened its mandate and rendered its landscape of agencies, bodies, funds, and programs more complex and interconnected. The UN deals with a wide variety of issue areas and is one of the most accessible IOs for nonstate actors (Tallberg et al. Reference Tallberg, Sommerer, Squatrito and Jönsson2013). In recent years, the organization has experienced criticism for its inability to act in various conflicts, as well as protests in association with its climate change summits (Gregoratti and Uhlin Reference Gregoratti, Uhlin, Tallberg, Bäckstrand and Scholte2018). The IO’s most powerful body, the UNSC, has relatively low legitimacy among the broader membership (Binder and Heupel Reference Binder and Heupel2015). Reform aimed at improving effectiveness and legitimacy is a consistent theme in the UN (Weiss Reference Weiss2012).

To examine popular legitimacy beliefs in the UN, we focus on confidence in the organization among citizens in all countries covered by the EVS and WVS 1994–2020, as well as the specific trajectories for a select number of countries: Germany, Japan, Poland, Russia, Turkey, and the US.Footnote ¹ These countries represent a mix between established powers with privileged positions on the UNSC (Russia and the US), regional powers without the same institutional privileges (Germany and Japan), and countries that have often been at the receiving end of global governance (Poland and Turkey).

The key finding in Figure 2.1 is that the perceived legitimacy of the UN among citizens declined between 1994 and 2014 in the observed countries, but has increased since. We see a decline of about 10 percentage points in the share of citizens having a great deal or quite a lot of confidence in the UN over the first two decades, from about 58 percent in 1994 to about 46 percent in 2014. Since 2015, this share has increased slightly to about 48 percent. However, we see great variation across countries.

Figure 2.1 Citizen confidence in the UN and domestic government

Source: EVS and WVS trend file, 3rd–7th wave, 1994–2020 (EVS 2021; Haerpfer et al. Reference Haerpfer, Inglehart and Moreno2021). Question wording: I am going to name a number of organizations. For each one, could you tell me how much confidence you have in them: is it “a great deal of confidence” (4), “quite a lot of confidence” (3), “not very much confidence” (2), or “none at all” (1). Number of respondents is ca. 1,000–2,000 per country. Poststratification weights were used. UN average calculated for all countries available. See www.worldvaluessurvey.org.

In Germany, slightly less than every other citizen has a great deal or quite a lot of confidence in the UN, which is slightly below the global average. Germany has seen much fluctuation in UN confidence since the early 1990s, in line with broader assessments that we are not seeing a continuous erosion of legitimacy, but rather short- or medium-term fluctuations (Kriesi Reference Kriesi2013; Walter Reference Walter2021). Between 1999 and 2004, UN confidence in Germany was at an all-time high during the observed period, perhaps influenced by the events around “9/11” and the ensuing wars in Afghanistan and Iraq, which were heavily politicized in Germany and broadly seen as underlining the need for multilateralism rather than unilateralism.

Turning to the patterns in other countries, Poland and the US stand out. In both countries, a growing share of citizens have confidence in the UN since 2010, despite both countries’ governments repeatedly expressing critique against the liberal world order. In Poland, UN confidence has risen from almost 50 to about 60 percent over this time period. In the US, it has risen from about 35–45 percent during the same time period. By contrast, in Russia and Turkey, UN confidence has fallen and is relatively weak. In Japan, UN confidence has remained largely stable at around 60 percent since the early 1990s.

One way of putting these patterns of IO confidence in a broader context is to compare with corresponding levels of citizen confidence in the respective domestic governments. Governments have on average not experienced the same decline as the UN over time in the observed countries, but start from lower levels of legitimacy in the eyes of citizens to begin with. The average level of confidence in governments is only marginally lower at the end of the observation period (41 percent) compared to the beginning (43 percent). Average confidence in the national government in the surveyed countries is consistently lower than the average confidence in the UN in the same countries. Today, every other citizen has confidence in the UN, while only 41 percent have confidence in their country’s government.

There are country-specific trends when comparing government and IO confidence. For example, in Japan, the gap between government and UN confidence was largest between 1994 and 2009 (almost 40 percent). Since then, confidence in the government has increased, shrinking the gap to about 10 percent today. In Russia, confidence in the government increased strongly from 30 to 50 percent over the past thirty years, while confidence in the UN fell slightly from 50 to about 45 percent over this period. Likewise, Poland has seen a surge of confidence in the government from about 50 to 70 percent during this period, while UN confidence fell from about 50 to slightly under 40 percent. In Germany, the trajectories for government and UN confidence are relatively similar. Finally, in the US, confidence in government has remained relatively stable at slightly under 40 percent since 1999, while confidence in the UN fell sharply between 1999 and 2009 from about 55 percent to about 35 percent.

Next, we turn to the regional organizations. We begin with the AU. This organization is a multi-issue organization active in a large number of policy areas. Central issue areas are development, economic integration, and security. The transformation of the Organization of African Unity (OAU) into the AU in 2001 brought about a new institutional structure, as well as deep policy reforms. More emphasis was placed on economic integration, environmental cooperation, and infrastructure development within the New Partnership for Africa’s Development (NEPAD). In addition, the AU’s pillar for peacekeeping was strengthened, by moving beyond the traditional principle of nonintervention to a pledge to human security, solidarity, and a responsibility to protect, captured in the African Peace and Security Architecture (APSA). As a main venue for African diplomacy and regional politics, the AU has assumed ever greater authority since the beginning of the 1990s (Hooghe et al. Reference Hooghe, Marks, Lenz, Bezuijen, Ceka and Derderyan2017) and has thus been increasingly judged against standards of democratic and effective decision-making (Nujoma Reference Nujoma2002; Witt Reference Witt, Dingwerth, Witt, Lehmann, Reichel and Weise2019). We examine perceptions of the AU in the context of six member countries with varying size, colonial past, and experiences with IOs: Kenya, Mali, Mozambique, Nigeria, South Africa, and Tanzania.

Average perceptions of how well the AU does its job are remarkably stable (Figure 2.2). The average level oscillates between 50 and 60 percent over the period from 2002 to 2015. Among the specific countries examined, AU perceptions fluctuate especially in Nigeria and Tanzania, where confidence dropped from 2008/2009 to 2012/2013, but then saw a recovery between 2012/2013 and 2014/2015. AU perceptions fluctuate in Mali and Mozambique as well, starting out with relatively high shares of people thinking the AU is helpful, but then experiencing a decline from 2008/2009 to 2012/2013, followed by a rise again in 2014/2015. By contrast, perceptions of the AU are relatively stable in Kenya and South Africa, where about every other citizen thinks that the AU is helpful.

Figure 2.2 Citizen support for the AU

Source: Afrobarometer, 2002–2015. Question wording: In your opinion, how much does the African Union do to help this country, or haven’t you heard enough to say? 0 = Do nothing, no help, 1 = Help a little bit, 2 = Help somewhat, 3 = Help a lot. Poststratification (“within-country”) weights were used. AU average calculated using data for all countries available, that is, eight countries before 2008, and about eighteen countries in Africa after 2008.

The EU is arguably the most powerful of all regional organizations. It enjoys more authority than other regional IOs because of decision-making competence in a larger number of issue areas, far-reaching delegation to supranational institutions, and extensive pooling of authority in collective decision-making (Hooghe et al. Reference Hooghe, Marks, Lenz, Bezuijen, Ceka and Derderyan2017; Zürn Reference Zürn2018). European integration is also deeper because of the monetary union between nineteen of the twenty-seven member states, and due to the primacy and direct applicability of EU law. Because of its far-reaching authority, the EU has become increasingly politicized in its member states (Hooghe and Marks Reference Hooghe and Marks2009; Zürn et al. Reference Zürn, Binder and Ecker-Ehrhardt2012; De Vries Reference De Vries2018; Schmidt Reference Dür, Marshall and Bernhagen2019). In January 2020, the UK left the EU following a national referendum on membership in 2016.

We analyze trust in the EU from 2003 to 2017. Average trust stood at about 45 percent in 2017, which is lower than the 60 percent recorded in 2007, but higher than the all-time low in 2013 at around 38 percent. We also show specific trends for Denmark, France, Germany, Italy, Poland, and the UK (Figure 2.3). These countries have varying experiences with the EU, in terms of when they joined the organization, the formal power they enjoy within institutions, and their contributions to the EU’s budget, which is an important dividing line in EU politics (Hooghe and Marks Reference Hooghe and Marks2005).

Figure 2.3 Citizen trust in the EU

Source: Standard Eurobarometer, 2003–2017. Question wording: I would like to ask you a question about how much trust you have in certain institutions. For each of the following institutions, please tell me if you ‘tend to trust’ it (1) or ‘tend not to trust’ it (0). Number of respondents is ca. 1000–2000 per country covered. Poststratification weights were used. The EU average is based on data from the EU member states. See.

In most of these countries, EU trust declined following the outbreak of the global financial crisis in 2007/2008, leading to historically low trust levels in 2013. However, EU trust has since recovered in most of the examined countries (Denmark, France, Germany, Italy, and the UK), the exception being Poland. EU trust is highest in Poland (and Denmark) among the countries studied, but has steadily declined from almost 80 percent in 2007 (following the country’s accession to the EU and its improving economic situation) to about 50 percent since 2013. Unsurprisingly, the lowest levels of trust throughout the time period are found in the UK, where trust fluctuates between 20 and 40 percent.

We now turn to Mercosur, a regional trade arrangement between Argentina, Brazil, Paraguay, Uruguay, and Venezuela.Footnote ² The organization is a customs union with free trade among its member states and common external tariffs. In addition, it coordinates common standards for goods, agriculture, health, infrastructure, and food. With its main focus on economic integration, Mercosur is more limited in terms of authority and issue scope than the AU and the EU. Moreover, its three main decision-making bodies – the Common Market Council, Common Market Group, and the Trade Commission – use unanimous decision-making. Mercosur has enacted a series of reforms to spur integration over the past two decades, including the deepening of judicial integration through the creation of the Permanent Review Court (Arnold and Rittberger Reference Arnold and Rittberger2013). However, Mercosur’s authority remains relatively restricted, both in terms of delegation and pooling (Hooghe and Marks Reference Hooghe and Marks2015; Meissner Reference Meissner and Krapohl2017). We focus on average perceptions of Mercosur in all member states, as well as the specific trajectories for three member countries: Argentina, Brazil, and Venezuela.

On average, in 2014–2015, almost 70 percent of all citizens in Mercosur evaluate the organization as 6 or higher on a 10-point scale ranging from very bad to very good. This is the highest level of average public trust among the IOs examined in this chapter. From 2000 to 2015, average positive evaluations of Mercosur rose from slightly less than 50 percent to about 67 percent. The rise was especially steep between 2009 and 2015, when average support increased more than 10 percentage points. The trajectories among the three member countries specifically examined are remarkably close to the average, in particular since 2003 (Figure 2.4).

Figure 2.4 Citizen support for the Mercosur

Source: Latinobarometer, 2001–2015. Question wording: From the list of institutions which are on this card, please put the institution on a scale from 1 to 10, 1 being very bad and 10 very good. Percentage of those scoring 6 or higher on this scale. Poststratification weights were used. Mercosur average calculated using data for all five member states. Data available for 2001, 2002, 2003, 2009, 2013, and 2015.

As a last step, we broaden the selection of IOs and use data from the most recent wave of the WVS (Figure 2.5). On the whole, citizens have a medium level of confidence in IOs. The share of citizens having a great deal or quite a lot of confidence in an IO is highest for the WHO (about 60 percent) and the UN (about 50 percent), followed by the WTO, the World Bank, and the ICC (40–45 percent). The corresponding figure for the IMF and NATO is about or slightly less than 40 percent. Overall, these figures suggest that IOs engaged in human security governance (ICC, NATO, UN, and WHO) enjoy greater confidence on average than IOs active in economic governance (IMF, World Bank, and WTO). Interestingly, confidence in domestic government is at slightly over 40 percent, putting national governments in the middle of the ranking.

Figure 2.5 Citizen confidence in IOs and domestic government

Source: WVS7, 2017–2020. Question wording: I am going to name a number of organizations. For each one, could you tell me how much confidence you have in them: is it “a great deal of confidence” (4), “quite a lot of confidence” (3), “not very much confidence” (2), or “none at all” (1). Number of respondents is ca. 1000–2000 per country. Poststratification weights were used. UN average calculated for all countries available, i.e., ca. 70 countries. See www.worldvaluessurvey.org.

Taken together, there is no consistent pattern indicating a crisis of legitimacy for IOs, even if there is a downward trend for the UN and the EU in a long-term perspective. However, both the UN and the EU have seen increasing legitimacy since about 2013–2014. In addition, confidence in the AU has been relatively stable over the past twenty years, while Mercosur has seen a significant rise in popular legitimacy. When comparing confidence across IOs, the WHO, and the UN enjoy most confidence.

Data

Few surveys have sought to capture elite opinion toward IOs, and existing datasets cannot be used for cross-national and cross-organizational comparisons of attitudes toward IOs among political and societal elites. Existing studies have drawn on data on elite opinion toward the EU (Hooghe Reference Hooghe2002; Best et al. Reference Best, Lengyel and Verzichelli2012; Persson et al. Reference Persson, Parker and Widmalm2019), and legitimacy beliefs among elites in government (Binder and Heupel Reference Binder and Heupel2015) as well as civil society (Gregoratti and Uhlin Reference Gregoratti, Uhlin, Tallberg, Bäckstrand and Scholte2018). The Chicago Council on Global Affairs regularly maps the opinions of both the general public and opinion leaders on matters of foreign policy, but its coverage is restricted to the US and it does not specifically focus on attitudes toward IOs (Kafura Reference Kafura2019).

In the absence of systematic data on elite opinion toward IOs, our team in the LegGov research program engaged in a large-scale collection of original data on elite legitimacy beliefs. The methodology is described in detail in Verhaegen et al. (Reference Verhaegen, Dellmuth, Scholte and Tallberg2019) (for in-depth analyses based on these data, see Tallberg and Verhaegen Reference Tallberg and Verhaegen2020; Scholte et al. Reference Scholte, Verhaegen and Tallberg2021; Verhaegen et al. Reference Verhaegen, Aart, Scholte and Tallberg2021; Dellmuth et al. Reference Dellmuth, Scholte, Tallberg and Verhaegen2022a, Reference Dellmuth, Scholte, Tallberg and Verhaegen2022b). Our survey was fielded in 2017–2019 in six diverse countries around the world – Brazil, Germany, the Philippines, Russia, South Africa, and the US. We also included a sample of global elites, working for transnational organizations, such as staff of IOs, multinational corporations, global news media, and international NGOs. More specifically, we interviewed 860 elite individuals: 124 in Brazil, 123 in Germany, 122 in the Philippines, 108 in Russia, 123 in South Africa, 122 in the US, and 138 at the global level.

We define elites as persons who hold leading positions in key organizations in society that strive to be politically influential. While most studies of elite opinion focus exclusively on political elites, the LegGov survey also encompasses wider societal elites. Specifically, we distinguished between six elite types, of which two types were political – party and bureaucratic elites – and four types were societal – business, civil society, media, and academic elites. In the global sample, we distinguish between member state representatives and IO officials in the category of international bureaucrats. For the selection of elite interviewees, the survey used quota sampling as the preferred procedure.

Among many issues, the survey inquired about elite individuals’ confidence in ten IOs. These IOs differ with respect to authority, membership, and issue area. They include IOs in economic and human security governance, similarly to the WVS data above; in addition, the elite data allow for distinguishing between global organizations with universal membership – the ICC, IMF, UN, UNFCCC, WHO, World Bank, and WTO – and restricted membership – the Group of Twenty (G20), NATO, and UNSC.

The elites who participated in the survey have higher formal education than the wider public. In addition, they likely have greater experience of interaction with IOs, as a result of working in positions and sectors that bring them into closer contact with political institutions. As can be expected, a comparative analysis of elite and citizen knowledge in the five countries shows that elites tend to know more about global governance than citizens. We find that about 62 percent of the elite respondents in Brazil, Germany, the Philippines, Russia, and the US correctly answered three knowledge questions about the UNSC, the IMF, and Amnesty International, compared to about 13 percent of the citizens in these countries in the WVS7 (Dellmuth et al. Reference Dellmuth, Scholte, Tallberg and Verhaegen2022b).Footnote ³

In the following, we present patterns in elite legitimacy beliefs across IOs, countries, and elite types.

Patterns in Elite Legitimacy Beliefs

On average, elites are moderately supportive of IOs. In contrast to the citizen data in Figure 2.5, a slight majority of elites have a great deal or quite a lot of confidence in these IOs (Figure 2.6). The percentage of elites having a great deal or quite a lot of confidence is highest in the WHO (more than 80 percent) and lowest in the G20 (about 50 percent). This result for the WHO is interesting given the growing contestation of the WHO during the COVID-19 pandemic in 2020. Like in the citizen data, the IO enjoying the second-highest level of support is the UN, with about 70 percent of elites having a great deal or quite a lot of confidence in the organization, followed by the ICC and the UNFCCC with slightly more than 60 percent. The World Bank, WTO, IMF, UNSC, and NATO all attract a great deal or quite a lot of confidence among slightly more than 50 percent of the elites studied.

Figure 2.6 Elite confidence in IOs and domestic government

Source: LegGov Elite Survey (Verhaegen et al. Reference Verhaegen, Dellmuth, Scholte and Tallberg2019), 2017–2019. Question wording as in the WVS7, with four answer categories (see Figure 2.1).

However, these averages hide important divisions among elites in terms of their levels of confidence in IOs. If 60 percent of elites have a great deal or quite a lot of confidence in IOs, 40 percent of elites do not. These contrasting perspectives on the legitimacy of IOs can help us to understand why elites compete in shaping citizen opinion toward IOs using positive and negative communication. The divisions among elites are particularly pronounced with regard to the IMF, NATO, and UNFCCC, for which a larger share of respondents use the two extreme options – a great deal of confidence or no confidence at all. The divisions are less severe in the case of the UN, WTO, and World Bank, for which a larger share of elite respondents use the two middle options – little or quite a lot of confidence in the IO.

Again, we use confidence in the domestic government as a point for comparison. Most IOs examined enjoy more elite confidence on average than the domestic governments, except for the IMF, UNSC, NATO, and G20. This might be because these IOs are especially contested in some of the examined countries, which is what we can observe in the next graph.

Figure 2.7 shows how confidence in the same ten IOs (pooled) varies across countries. Interestingly, elites on average have most confidence in IOs in Brazil, the global sample, the US, and Germany, and least in the Philippines, Russia, and South Africa. Only in South Africa is the support for IOs among elites clearly lower than in the other countries; in South Africa, only slightly more than 40 percent have a great deal or quite a lot of confidence in IOs, whereas the corresponding figures for the other samples are between about 55 and 70 percent.

Figure 2.7 Elite confidence in IOs, by country

Source: LegGov Elite Survey (Verhaegen et al. Reference Verhaegen, Dellmuth, Scholte and Tallberg2019), 2017–2019. Question wording as in the WVS7, with four answer categories (see Figure 2.1). This figure is based on the average percentage of people having a great deal or quite a lot of confidence in the ten IOs shown in Figure 2.6.

At the same time, elites in some of these countries have been at the forefront of attacking international cooperation. For example, in the US, former President Donald Trump concretized his “America first” strategy by withdrawing the country from several international agreements and organizations; in the Philippines, President Duterte is a vocal antagonist of international cooperation and has threatened to leave the UN; in Germany, the populist right-wing party Alternative für Deutschland is critical of EU cooperation and has expressed support for Brexit; and in Brazil, President Bolsonaro regularly offers scathing critique of the UNFCCC and the WHO. Examining the variation in confidence in all IOs in these countries, we see that elites are especially divided over international cooperation in Brazil, the Philippines, and the global sample, and least divided in Germany, Russia, and the US. While this may seem counterintuitive given partisan political polarization in the US, it is worth reiterating that the partisan-political category only is one of six different elite types in our sample.

We now turn to the question of how confidence in IOs varies by elite type. Figure 2.8 shows that there is fairly extensive variation across the six types of elites, amounting to almost 30 percentage points between the lowest and the highest confidence levels. In particular, bureaucrats in various categories (national bureaucrats, IO officials, and member state representatives in IOs) and business leaders stand out as having high and very similar levels of confidence toward IOs. Around 70 percent of elites in these sectors have a great deal or quite a lot of confidence in IOs. In the categories of party and research elites, around 60 percent have a great deal or quite a lot of confidence in IOs. The lowest shares of elite individuals having a great deal or quite a lot of confidence in IOs can be found among media and civil society elites (around 50 percent).

Figure 2.8 Elite confidence in IOs, by elite type

Source: LegGov Elite Survey (Verhaegen et al. Reference Verhaegen, Dellmuth, Scholte and Tallberg2019), 2017–2019. Question wording as in the WVS7, with four answer categories (see Figure 2.1). Member state representatives in IOs (“MS rep’s”) as well as permanent officials in IOs (“IO officials”) are categories which belong to the global sample depicted in Figure 2.6, while the other categories consist of data pooled from the six countries examined. This figure is based on the average percentage of people having a great deal or quite a lot of confidence in the ten IOs shown in Figure 2.6.

In terms of how divided elites are, we see a larger share of respondents indicating a great deal of confidence or no confidence at all among bureaucrats, member state representatives in IOs, and media. The divisions are less severe among research elites, business elites, and IO officials.

In sum, elites show moderate support for IOs on average, even if there is some variation across IOs, countries, and elite types. Elites have most confidence in the WHO and the UN, and least confidence in NATO and the G20. The highest confidence in IOs is found in Brazil and the global sample, and the lowest in South Africa and Russia. Bureaucratic elites and business elites display the highest confidence in IOs, while confidence is lowest among civil society and media elites.

However, these averages also mask divisions among elites within the same country and sector. If we consider the variety of responses given by elites, it becomes clear that they are split in their attitudes toward IOs. Those divisions are particularly strong with regard to the IMF, NATO, and the UNFCCC; in Brazil, the Philippines, and the global sample; and among bureaucrats, media elites, and member state representatives in IOs. In the next section, we examine how elites with varying attitudes toward international cooperation make use of positive and negative messages in an effort to influence the public.

Data

Legitimation and delegitimation of IOs may be of three principal kinds: discursive, institutional, and behavioral (Bäckstrand and Söderbaum Reference Bäckstrand, Söderbaum, Tallberg, Bäckstrand and Scholte2018; Tallberg and Zürn Reference Tallberg and Zürn2019). Given our focus on elite communication, we are exclusively interested in discursive legitimation, as expressed in public statements and official texts by elite supporters and opponents of IOs (Steffek Reference Steffek2003; Gronau Reference Gronau2015). Discursive legitimation and delegitimation takes place in a variety of ways, among them, official IO communication, parliamentary debates, press conferences, party programs, news media, and social media. Different forms of content analysis are the favored method for capturing discursive legitimation and delegitimation.

The existing literature on discursive legitimation and delegitimation is sizeable and until recently mainly case-study based. Many studies focus on the EU (for overviews, see Hurrelmann Reference Hurrelmann, Schneider, Nullmeier and Hurrelmann2007; Rittberger and Schroeder Reference Rittberger, Schroeder, Börzel and Risse2016), but a growing number of contributions study other IOs, often comparatively. For instance, Dingwerth et al. (Reference Dingwerth, Witt, Lehmann, Reichel and Weise2019) use IO annual reports from the period 1970–2010 to examine the increasing emphasis on democratic procedures in IOs’ public communication, focusing on the AU, International Atomic Energy Agency, International Union for Conservation of Nature, United Nations High Commissioner for Human Rights, and WTO. Gronau (Reference Gronau2015) studies the self-legitimation efforts of the G8 and the G20 over almost 40 years based on ample textual and visual material. Nuñez-Mietz (Reference Nuñez-Mietz2018) evaluates the case of NATO’s intervention in Kosovo in 1999, identifying seven types of strategic self-legitimation moves. Gregoratti and Uhlin (Reference Gregoratti, Uhlin, Tallberg, Bäckstrand and Scholte2018) study protests by civil society activists targeting global governance institutions, using illustrations from the ASEAN, Asian Development Bank, and G8. Bexell et al. (Reference Bexell, Jönsson and Uhlin2022) offers paired comparisons of the legitimation and delegitimation of global governance institutions in a variety of issue areas.

The most recent years have seen a number of comparative large-scale data collection efforts on legitimation and delegitimation. These initiatives have typically relied on content analysis of either IO annual reports (e.g., Dingwerth et al. Reference Dingwerth, Witt, Lehmann, Reichel and Weise2019; Gregoratti and Stappert Reference Gregoratti and Stappert2019; Bexell et al. Reference Bexell, Jönsson and Stappert2021, Reference Bexell, Jönsson and Uhlin2022; Lenz et al. Reference Lenz, Schmidtke, Krösche and Schirmer2020) or news media (e.g., Bes et al. Reference Bes, Sommerer and Agné2019; Schmidtke Reference Schmidtke2019; Rauh and Zürn Reference Rocabert, Schimmelfennig, Crasnic and Winzen2020; Sommerer et al. Reference Sommerer, Agné, Zelli and Bart2022). The aim of these studies is to provide general knowledge about legitimation and delegitimation by comparing such practices across organizations and over time. Social media appear to offer a source of data on discursive legitimation of IOs that so far has remained largely untapped.

Patterns in Legitimation and Delegitimation

We discuss legitimation and delegitimation in terms of three dimensions: intensity, narratives, and tone (Tallberg and Zürn Reference Tallberg and Zürn2019). Intensity refers to the number of legitimation or delegitimation events (e.g., statements) within a given time frame, sometimes referred to as the level of “politicization” (e.g., De Wilde and Zürn Reference De Wilde and Zürn2012; Rauh and Zürn Reference Rocabert, Schimmelfennig, Crasnic and Winzen2020). Narratives are patterns in the standards invoked to justify or challenge IOs. Tone captures whether discursive practices frame the IO in positive terms (legitimation) or negative terms (delegitimation). In the following, we use these three dimensions to summarize what existing empirical literature can tell us about discursive legitimation and delegitimation of IOs.

Intensity. Overall, existing research provides a relatively unanimous picture of the intensity of legitimation and delegitimation over time. IOs have seen an overall increase in the intensity of legitimation and delegitimation since the 1980s until today. However, this development has not been uniform over time or across IOs. IOs saw more intense legitimation and delegitimation struggles from the early 1980s until the mid-2000s, after which intensity has declined and the trend mainly involves minor fluctuations. Since the mid-2000s, patterns in intensity appear to be more IO-specific, with some IOs experiencing overall increases in the intensity of legitimation and delegitimation, while other IOs experience peaks around specific events, such as the global financial crisis 2007–2008.

To discuss the evidence in more detail, Schmidtke (Reference Schmidtke2019), for instance, comparatively examines intensity in legitimation and delegitimation discourse about the EU, G8, and UN in the quality press of four countries – Germany, Switzerland, the UK, and the US – over the time period 1998–2013. The findings reveal that there is no clear trend of growing contestation over time. Rather, we see rising public discourse about all three institutions from the late 1990s until the mid-2000s, followed by a marked decline in intensity. While all three IOs largely follow the same pattern, the legitimacy discourse around the EU is almost always more intense than the discourse about the G8 and UN. Schmidtke attributes this pattern to variation in the authority of the three organizations.

Lenz et al. (Reference Lenz, Schmidtke, Krösche and Schirmer2020) embark on a comparative inquiry into legitimation discourse across a large number of regional organizations over time. Drawing on the annual reports of twenty-seven organizations between 1980 and 2019, they arrive at several key findings. Legitimation intensity in the context of the IOs in their sample has risen over the observation period, reaching a plateau around 2005, after which average legitimation intensity has remained relatively stable or even declined slightly. However, when they group IOs by world regions, they find that developments in legitimation intensity vary considerably. While IOs in Africa and the Americas follow the general pattern, IOs in Europe see a slight continuous increase over time, and IOs in the Asia-Pacific see a U-shaped pattern with declining legitimation intensity until around 2000 followed by an increase ever since.

Rauh and Zürn (2020) study legitimation and delegitimation in the context of the IMF, NAFTA, World Bank, and WTO based on a semiautomated analysis of statements in more than 120,000 articles in international newspapers during the period 1992–2012. They confirm that the surge in IO authority over the past thirty years has paralleled a trend toward more intense (and differentiated) legitimation and delegitimation discourse. They show that the share of legitimation and delegitimation statements about the IMF and the World Bank in their newspaper corpus increased from 1992 to 2005, after which it declined slightly and then has fluctuated. The patterns for NAFTA and the WTO are more fluctuating overall.

Dingwerth et al. (Reference Dingwerth, Schmidtke and Weise2020) also observe an increase in legitimation intensity, measured as the share of identity- and purpose-related statements in the annual reports of twenty cross-regional IOs. This increase in the use of legitimation statements by IOs roughly follows a linear trend over the decades in their observation period, from 1980 to 2011. Over time, the share of legitimation statements increased more than twofold, from 12 percent in 1980 to as much as 28 percent in 2017.

Finally, Sommerer et al. (Reference Sommerer, Agné, Zelli and Bart2022) map the intensity of delegitimation for thirty-two IOs over the time period 1985–2020, based on media coverage of public criticism against IOs, in a larger effort to identify legitimacy crises. They find that contestation of IOs increased from the 1980s onwards to reach a peak around the turn of the millennium. Some IOs subsequently experienced rising intensity in delegitimation in association with the global financial crisis, while other IOs saw such increases in the most recent years. Overall, however, IOs tend to follow unique patterns of contestation, rather than conform to a uniform picture of growing delegitimation, indicative of a legitimacy crisis in their analysis.

Narratives. Generally, existing research indicates that the type of narratives used to legitimize and delegitimize IOs have broadened over time. There appears to be a relation between increased authority of IOs and greater variety in the narratives used to justify this authority. Purpose- and performance-related narratives are typically invoked more often than procedure-related narratives in IOs’ own communication, but the latter type is on the rise. In particular, IOs increasingly legitimize themselves by referring to democratic values. Moreover, the types of narratives used to criticize and defend IOs vary to some extent: While actors that aim to delegitimize IOs typically invoke fairness and democracy, actors that seek to legitimize IOs invoke a broader range of standards pertaining to procedure, performance, and purpose.

To review the evidence in greater detail, Zürn (Reference Zürn2018, 70–77), for instance, distinguishes between seven different legitimation narratives: participatory, legal, fairness, technocratic, traditional, relative gains, and manipulative narratives. One central conclusion of his work is that most IOs rely on technocratic narratives to justify their exercise of authority (see also Uhlin Reference Uhlin2019). However, the level of authority that is exercised by many IOs makes it necessary to complement justifications on technocratic grounds with other types of narratives, especially in times of politicization. Examples are fairness- and participation-related narratives, which are found to expand in times of more intense public debate and protest. Thus, as IOs gain increasing authority, we observe a broadening of the narratives used to justify this authority.

Lenz et al. (Reference Lenz, Schmidtke, Krösche and Schirmer2020) distinguish between procedure-, performance-, and purpose-related narratives about IOs, which they then characterize in terms of liberalism, communitarianism, and functionalism, yielding nine legitimation narratives. They arrive at several important findings based on their analysis of communication from twenty-seven IOs over the time period 1980–2019. All nine types of narratives are present in IO legitimation communication. This suggests that IOs make use of a broad range of standards when justifying their authority. However, functional narratives are vastly more common than liberal or communitarian narratives. Likewise, purpose-based narratives are more common than performance-based narratives and especially procedure-based narratives. These patterns are largely robust across IOs in different world regions. However, if we consider patterns over time in the development of these narratives, the evidence yields a more varied picture, highlighting alternative temporal paths.

Rauh and Zürn (2020) pit technocratic, fairness-based, and participatory narratives against each other. They demonstrate increasing shares of civil society organizations (CSOs) in legitimation narratives until 2006, and declining rates since, which indicates that CSO presence in legitimation narratives might be related to legitimation intensity more broadly. In addition, they show that civil society actors, when making statements about IOs in news media, invoke fairness-based demands that stretch beyond the technocratic legitimation narrative that conventionally characterizes global economic governance.

Dingwerth et al. (Reference Dingwerth, Witt, Lehmann, Reichel and Weise2019) focus specifically on democracy-related narratives about IOs. Their case studies convey two important findings. First, “the people” is becoming a central reference point in the legitimation of the five IOs they study. IOs are increasingly asked to demonstrate not only what they can do for their member states but also for the citizens in these states. Second, procedural legitimacy standards are gaining ground, as IOs are increasingly evaluated on the basis of how they make decisions and not only what they accomplish. Dingwerth et al. (Reference Dingwerth, Witt, Lehmann, Reichel and Weise2019) point to growing politicization and a rise in nonstate actors making demands on IOs as explanations for these trends. In a related study, Dingwerth et al (Reference Dingwerth, Schmidtke and Weise2020) analyze data on the democratic legitimation of twenty regional organizations in annual reports from 1980 to 2011. They demonstrate a far-reaching rise of democratic legitimation in global governance and find that politicization emerges as the main driving force behind a “democratic turn” in IO legitimation.

Bexell et al. (Reference Bexell, Jönsson and Uhlin2022) examine the narratives that are present in the legitimation and delegitimation of sixteen global governance institutions in a variety of issue areas. While a central conclusion is that narratives tend to be quite context dependent, they also identify a number of larger patterns. Notably, actors engaging in legitimation of IOs tend to invoke technocratic norms of expertise and effectiveness, democratic norms of transparency and participation, and norms linked to the specific purpose of the organization. When actors instead seek to delegitimize IOs, they are more likely to invoke norms related to fairness, but also democratic standards.

Other types of narratives relate to more specific empirical story lines. For instance, narratives can invoke mandates, as in the case of the Arab League’s backing and the AU’s condemnation of the Libya intervention in 2011 (Drieskens and Reykers Reference Drieskens, Reykers, Biermann and Koops2017). Other criticism that IOs have had to fend off in the past revolve around their memberships, institutional structures, and unique political challenges (Zaum Reference Zaum2013). For instance, US politicians have questioned the legitimacy of NATO’s unanimity rule that legally accords the US a voice in alliance decision-making equal to Lithuania’s (Michel Reference Michel and Mayer2014). Finally, elite communication sometimes pertains to the norms and rules that an IO stands for. For example, the EU’s high standards for minority rights face strong criticism because several member states themselves do not meet these standards (Gawrich Reference Gawrich2006).

Tone. Previous research seeking to unravel patterns of legitimation and delegitimation shows that elites most often communicate about IOs in negative terms, except in the context of self-legitimation. However, we need more comparative evidence to be able to make conclusive statements about the tone of communication in global governance. Based on available evidence, it appears that the UN is more often subject to positive communication than the EU and economic organizations such as the IMF, but that specific organizations within the UN system, such as the UNSC, are predominantly delegitimized rather than legitimized.

Taking a closer look at the evidence, Schmidtke (Reference Schmidtke2019) sheds light on the tone of the legitimacy discourse around the EU, G8, and UN, mapping the share of positive evaluative statements about the three IOs. His analysis suggests that the tone in news media in Germany, Switzerland, the UK, and the US is predominantly negative. The UN is usually greeted with the highest share of positive statements, while the tone is slightly more negative toward the EU and most negative toward the G8. Overall, the analysis suggests that positive and negative evaluations do not follow a uniform trend over time but are characterized by country-specific patterns. Lighter colors indicate more positive sentiment, whereas darker colors indicate more negative sentiment. There is no clear trend that negative communication has consistently become more prevalent over time based on the evidence presented by Schmidtke (Reference Schmidtke2019).

Focusing on the IMF, Tokhi (Reference Tokhi, Stephen and Zürn2019) reveals that this organization is mainly talked about negatively in elite communication. Using data on all the statements made by member state representatives during the meetings of the International Monetary and Financial Committee in 2010 and 2014, Tokhi (Reference Tokhi, Stephen and Zürn2019) shows that more than 60 percent of all statements contested the status quo. By contrast, positive evaluations make up about one-third of all statements. However, the intensity with which some member states challenged or embraced the status quo varied considerably across constituencies. There is striking country-specific variation, where the industrialized democracies display larger proportions of positive statements and the rising powers, especially the BRICS, primarily make contesting statements.

Binder and Heupel (Reference Binder and Heupel2015) assess legitimation and delegitimation of the UNSC in the UN General Assembly and find that the UNSC suffers from a “legitimacy deficit.” Negative statements about the UNSC, especially its procedures, outweigh positive evaluations. The evidence comes from a sample of seven debates in the General Assembly over the time period 1991–2009. Binder and Heupel (Reference Binder and Heupel2015) emphasize that, despite substantial delegitimation efforts by states demanding reforms, there is also evidence that states acknowledge when the UNSC seeks to legitimize itself by improving decision-making procedures or performing well.

Taken together, the reviewed patterns in legitimation discourse suggest that: (a) IOs saw more intense legitimation and delegitimation struggles from the early 1980s until the mid-2000s, after which intensity has declined and mainly involves minor fluctuations; (b) purpose- and performance-related narratives are generally invoked more often than procedure-related narratives in IOs’ own communication, but the latter type appears to be on the rise; and (c) IOs are mostly communicated about in negative messages, except in the context of self-legitimation.

Illustrations

These patterns beg important questions about how elites talk about IOs in public spaces. To shed light on this, we illustrate intensity, tone, and narratives in elite communication as present in varying media outlets. When communicating about IOs, elites are nowadays navigating a more complex media landscape, where traditional news media exist alongside novel social media as two main pathways for conveying messages about IOs.

For these purposes, we draw on news media (both print and online) and social media (Twitter) data collected within the LegGov research program for an analysis of legitimation around sixteen IOs. These data enable us to identify predominant narratives about specific IOs during a given time period, as well as the tone of these narratives (positive or negative). The time period chosen in our case is January to August 2020, which was the most recent time period available at the time of research. This period of six months provides us with sufficient material to illustrate how actors engage in legitimation and delegitimation using news media and social media. We identify narratives by using a method that captures frequent “associations” between words (i.e., words that co-occur most frequently with a selected word) and tone by conducting a sentiment analysis. While this method allows us to capture prominent narratives in news media and social media, it also reflects the broader context in which IOs are discussed at a certain point in time, which does not correspond to legitimation or delegitimation, strictly speaking.

The strategy to examine both news media and social media is novel for research about elite communication and legitimation (Hall et al. Reference Hall, Lenz and Obydenkova2021) and offers an opportunity to capture and compare how people talk about IOs in traditional and novel media outlets (Anstead and O’Loughlin, Reference Anstead and O’Loughlin2015). While these outlets are not reserved for political and societal elites exclusively, such elites likely dominate both news and social media communication about IOs. In the quality press journalists tend to grant preferential access to information and statements from high-ranking governmental actors and large businesses (Danielian and Page Reference Danielian and Page1994; Tresch Reference Tresch2009). On Twitter, governments, parties, NGOs, and IOs themselves are regular communicators about global governance; in addition, individual Twitter communicators are more likely to belong to elite groups, since they on average have higher levels of education and income than the general public (Pew Research Center 2019).

News media and Twitter data were collected with the assistance of Weblyzard Technology – an Austrian semantic technology company with a strong record of applied research in collaboration with universities and organizations such as UN Environment and BMW Group. Weblyzard’s software was used to systematically collect all references to our privileged IOs in online and print news articles as well as tweets during this six-month period (see Scharl et al. Reference Scharl, Hubmann-Haidvogel, Weichselbraun, Lang, Sabou and Sprague2013, Reference Scharl, Föls, Herring, Piccolo, Fernandez, Alani and Bagnoli2016, Reference Scharl, Herring, Rafelsberger, Hubmann-Haidvogel, Kamolov, Fischl, Föls and Weichselbraun2017). The data collection resulted in (a) tens of thousands of news media articles from 315 English-speaking online and print news sources, such as The Guardian, Canberra Times, Financial Times, Associated Press, and (b) 622,866 tweets.

For the analysis, we select four central IOs with extensive authority in their respective issue area – the EU, IMF, UNFCCC, and UNSC – and which occur in several of our experimental studies. We do separate analyses for these IOs and present patterns for each organization individually. We use this method to capture intensity by identifying the most frequent associations between the search term (the IO) and its strongest word associations (which we broadly interpret as narratives) at a specific point in time. The keyword graphs display a “semantic network,” which captures what negative and positive associations (tone) are present in communication about an IO.

For this purpose, we use a sentiment analysis, which classifies a news media article or a tweet into the categories negative or positive. The color of the data points captures the sentiment of a particularly prominent word association (or narrative). The analysis is performed by using dictionaries of about 12,000 a priori established English-language terms tapping either negative or positive emotions. Thus, our data are illustrative of topics that were intensively discussed during the period studied, and the extent to which these discussions were negative or positive in tone (see Weichselbraun et al. Reference Weichselbraun, Gindl, Fischer, Vakulenko and Scharl2017 for a detailed discussion of the methodology).Footnote ⁴ With regard to tone, it is worth reiterating that the sentiment analysis captures the tone of the contexts in which IOs are debated and not necessarily the intention of elites to legitimize or delegitimize IOs.

We proceed by discussing the results for each of the four selected IOs. We present the analysis in pairs (Figures 2.9–2.12), comparing narratives around the IO in news media (top panel) and social media (bottom panel). Lighter colors indicate more positive sentiment, whereas darker colors indicate more negative sentiment.

We start by discussing the EU (Figure 2.9). The news media data reveal that the EU during this period was mostly discussed in negative contexts, such as the presidential elections in Belarus in August 2020, the refugee camps in Greece and Turkey, and the development of the economy within the Eurozone. One salient narrative involved mostly positive associations, namely reporting on the EU’s Coronavirus recovery fund. In the same narrative, Dutch Prime Minister Mark Rutte was primarily discussed in negative terms as a proponent for countries resisting an expansive financial recovery package. In comparison, the Twitter discussion about the EU during this period mainly features positive narratives. The Coronavirus recovery fund did not play a similarly large role on Twitter. Rather, positive narratives focused on an agreement with the UK over Brexit and the future of trade relations between Britain and the EU after Brexit, both processes in which the EU plays a pivotal role as a negotiator. On the negative side, Twitter communicators discussed the risk of a trade war with China, which would negatively affect EU trade.

Figure 2.9 Legitimacy narratives in relation to the EU in news and social media

Notes: LegGov Newspaper and Twitter Dataset, January–August 2020. Lighter colors indicate more positive sentiment, whereas darker colors indicate more negative sentiment.

Next, we examine communication around the IMF (Figure 2.10). In news media, the IMF, just like the EU, is mostly debated in contexts that have negative associations. One example is the development of the Argentinian economy, which the IMF sought to support with a nearly $57 billion line of credit agreed with Argentina in the Stand-By-Agreement from 2018, but which failed to prevent a drastic depreciation of the Argentinian Peso since July 2019. On Twitter, the dominant narratives involving the IMF were mainly negative in tone and pertained to comparisons between the Coronavirus recession and the Great Depression, as well as Coronavirus-related lockdowns in Africa possibly calling for financial support by the IMF and the World Bank. Yet the IMF approval of an 18-month $5 Billion financial plan for Ukraine was mostly debated in positive terms.

Figure 2.10 Legitimacy narratives in relation to the IMF in news and social media

Notes: LegGov Newspaper and Twitter Dataset, January–August 2020. Lighter colors indicate more positive sentiment, whereas darker colors indicate more negative sentiment.

In the case of the UNFCCC (Figure 2.11), the dominant news media narratives during this period revolved around Greta Thunberg, the Paris Agreement, António Guterres, and the climate negotiation summit. The context in which Greta Thunberg is discussed is mostly negative, because news media report on her criticism of the performances of countries in the UNFCCC. On Twitter, the discussion around the UNFCCC is considerably more positive, even when focusing on similar topics, including Greta Thunberg, because Twitter communicators tend to praise these efforts. One of few negative narratives revolves around the delay of the COP26 UN Climate Change Conference, postponed due to the Corona pandemic.

Figure 2.11 Legitimacy narratives in relation to the UNFCCC in news and social media

Notes: LegGov Newspaper and Twitter Dataset, January–August 2020. Lighter colors indicate more positive sentiment, whereas darker colors indicate more negative sentiment.

In the context of the UNSC (Figure 2.12), the narratives in news media were again more critical in tone than on Twitter. In the news, salient narratives focused on topics such as the nuclear deal with Iran, the assassination of Iranian General Qasem Soleimani, Kenya being elected to the UNSC in competition with Djibouti, and Canada’s failed bid for a seat on the UNSC. Twitter discussions about the UNSC dealt with similar topics, but also the territorial dispute over the Kashmir region and the humanitarian crisis in Syria, in which the UNSC is involved by safeguarding aid delivery.

Figure 2.12 Legitimacy narratives in relation to the UNSC in news and social media

Notes: LegGov Newspaper and Twitter Dataset, January–August 2020. Lighter colors indicate more positive sentiment, whereas darker colors indicate more negative sentiment.

In sum, these illustrations exemplify how elite communication in media conveys both positively and negative framed narratives about IOs even if the latter tend to dominate, especially in traditional news media.