Introduction

In September 2011, José Manuel Barroso, then President of the European Commission, called the Eurozone financial crisis ‘the biggest challenge in the history’ of the European Union (EU). The crisis affected state budgets, the banking sector, and the real economy; it rattled the EU’s flagship project of Economic and Monetary Union (EMU) and, for a while, overshadowed most other EU activities. A series of emergency ‘bailouts’ for Eurozone states in danger of sovereign default, multiple interventions by the European Central Bank (ECB), and various reforms to EMU rules were needed until the Eurozone economy was stabilised. Yet, as Barroso pointed out, the crisis also had a more explicitly political component. ‘This crisis […] is also a crisis of confidence’, he told the European Parliament (EP) in his State of the Union Address. ‘A crisis of confidence in our leaders, in Europe itself, and in our capacity to find solutions’ (Barroso 2011).

The statement by Barroso highlights that the Eurozone crisis constituted challenge to the legitimacy of the EU, with the potential to undermine the social acceptance of its institutions (Longo and Murray 2015; Schweiger 2017). This legitimacy dimension of the crisis deserves attention because its effects might outlast the economic problems that caused the crisis. In the scholarly discussion, three main legitimacy concerns have been identified. First, the crisis undermined trust in the ability of EU institutions to produce good policy outputs. After all, the Eurozone’s troubles could be blamed in part on gaps in the institutional architecture of EMU (Krugman 2013; Scharpf 2013). Second, the crisis exposed certain undemocratic features of EU decision-making. The EU’s crisis governance was characterised by intergovernmental and technocratic policy modes, while parliaments at the EU and member-state level were largely sidelined (Crum 2013; White 2015). Third, the bailouts and the ECB interventions raised questions of legality, since they seemed to violate the spirit of treaty provisions that prevent the EU from assuming commitments of member-state governments and rule out ECB purchases of national debt instruments (Mayer 2012, pp. 103–116).

There has not been much empirical research, however, that examines whether these considerations have indeed resonated in the European public sphere. Did the EU’s legitimacy decline in the course of the Eurozone crisis? If so, how did this de-legitimation process unfold? Which political constellations and arguments fuelled de-legitimation? To answer these questions, this article examines political discourse in newspapers in four Eurozone member states—Germany, Austria, Spain, and Ireland—between 2009 and 2014. Our analysis proceeds in five steps. The first section discusses our conceptual framework and develops five hypotheses on how the Eurozone crisis affected the EU’s legitimacy. The second section explains the methods used in our analysis; it discusses the selection of countries and time periods as well as the technical steps taken in examining newspaper reporting. The following three sections present our empirical findings. We begin with a quantitative overview of legitimation trajectories over time, then analyse which discourse constellations (e.g. national contexts, speakers, and framing of newspaper statements) were most likely to produce de-legitimating statements, and finally perform a more detailed qualitative analysis of discourse in the time period that saw the greatest intensity of de-legitimation. In the concluding section, we discuss how our findings help us understand the effects of the Eurozone crisis on the EU’s legitimacy.

Conceptual considerations: from a policy problem to a legitimacy crisis

In empirical political research, the concept of legitimacy is used to describe a situation in which political rule is perceived as rightful by those subjected to it (Barker 1990; Gilley 2006). As Max Weber (1968, pp. 212–301) pointed out in his pioneering research on sources and types of legitimacy, the criteria that influence these perceptions vary across time and place. Legitimacy, as an empirical phenomenon, is not an inherent characteristic of political institutions or decisions; rather it is constructed in political activities and discourses (Suchman 1995; Barker 2001, pp. 21–29). Legitimacy is challenged if a political problem gives rise to activities and discourses in which the rightfulness of political rule is explicitly contested.

In the EU context, such legitimacy debates form a subset of the general EU-related contestation that emerges in the process of politicisation of European integration (de Wilde and Zürn 2012; Statham and Trenz 2013; Hutter et al. 2016). European integration is understood as politicised if EU affairs become a relevant topic in the member states’ political discourse (salience), trigger controversies (polarisation), and are debated by a broad range of societal actors (resonance/actor expansion) (de Wilde 2011; Hutter et al. 2016). However, politicisation does not necessarily imply that the EU’s legitimacy is being challenged; it may be restricted to policy aspects of European integration that do not raise issues of rightfulness of EU rule. This means that, in the analysis of contestation over EU affairs, the level of politicisation and the extent of legitimacy debates about EU rule should be considered as separate dimensions (Table 1). Four scenarios can be distinguished:

  • If politicisation levels are low and little attention is paid in EU-related debates to the rightfulness of EU rule, we have the situation described by Lindberg and Scheingold (1970) as permissive consensus on European integration. EU policy and EU legitimacy both remain largely undiscussed.

  • If politicisation levels are low but, in the EU-related debates that do occur, the rightfulness of EU rule becomes an issue, we are dealing with an untapped de-legitimation potential. This is the constellation described by Van der Eijk and Franklin (2004) in their influential metaphor of the EU as a ‘sleeping giant’—an entity ripe for politicisation that political elites are trying shield from public interest, for fear of undermining its legitimacy.

  • If politicisation levels are high but political discussions tend to focus on issues that do not raise questions about the rightfulness of EU rule, we are dealing with an ideal–typical policy debate. EU-related issues are salient and controversial, but the legitimacy of the EU rarely moves to the centre of attention.

  • Finally, if politicisation levels are high and the rightfulness of EU rule is controversially debated, we can speak of a legitimacy debate. Such a legitimacy debate need not result in an outright collapse of legitimacy. As some political actors challenge the EU’s legitimacy, others might jump to its defence. Indicators of a legitimacy crisis exist only if two conditions are met: (a) the legitimacy debate continues over extended time periods, suspending the normal ‘legitimation attention cycle’ (Hurrelmann et al. 2009, pp. 505–507) that sees legitimacy concerns being crowded out by discussions of new policy problems, and (b) assessments of the EU’s legitimacy become more negative as the debate persists.

These considerations are not only helpful in categorising various types of contestation about EU affairs, but also allow us to develop a hypothetical sequence that describes how policy problems like the Eurozone crisis may come to challenge the EU’s legitimacy.Footnote 1 The first step in this sequence—from permissive consensus to policy debate—consists of an intensification of debates about the policy in question. This step primarily entails an increase in politicisation. In a second step—from policy debate to legitimacy debate—we find an increased use of arguments that reach beyond policy and bring issues related to the rightfulness of the EU polity into play. (If the sequence starts from a situation of untapped de-legitimation potential, we can expect this shift to occur particularly quickly.) Finally, in a third step—from legitimacy debate to legitimacy crisis—discussions about the EU’s legitimacy persist even as the specific policy problem that triggered them recedes in public attention, and the tenor of these debates turns increasingly negative.

Table 1 Typology of EU-related political contestation
Full size table

In what follows, we use this sequence model to examine developments in EU-related political contestation about the Eurozone crisis. Our ambition is to find out (a) whether the Eurozone crisis has indeed led to an increase in discourse that explicitly challenges the legitimacy of the EU, and (b) which discourse constellations—that is, which configuration of political contexts, speakers, and arguments—are particularly likely to result in de-legitimation. In contrast to our inquiry, existing research on political contestation about the Eurozone crisis has focused primarily on the first step in the above sequence, the politicisation of EU governance. Research on a variety of discursive settings—parliamentary debates (Wendler 2016; Maatsch 2017), party manifestos (Hooghe and Marks 2018, pp. 117–118), media reporting (Kriesi and Grande 2016; Leupold 2016; Hutter and Kriesi 2019), as well as social media discourses (Michailidou 2017)—all demonstrate a spike in debates about EU governance in the crisis years; they also show that these debates remained elite-dominated and did little to constrain the EU’s technocratic crisis response (Schimmelfennig 2014; Börzel and Risse 2018). However, existing research has not systematically addressed the influence of the crisis on the EU’s legitimacy (for a qualitative case study of legitimation processes during the crisis, see Vaara 2014). Given the direct impact that the crisis had on many EU citizens, as well as its long duration, it is plausible to assume that crisis-induced political debates also had a legitimacy dimension and that the crisis did indeed push the EU in the direction of de-legitimation. This consideration results in our first two hypotheses.

H1

Legitimacy debates triggered by the Eurozone crisis become more prevalent over the course of the crisis.

H2

Legitimacy debates triggered by the Eurozone crisis turn more negative over the course of the crisis.

The crisis did, of course, have a differentiated effect on the various Eurozone states. Many researchers have emphasised the fault lines that emerged between the states that suffered large-scale banking and sovereign debt problems in their domestic economy (‘crisis states’) and others that primarily experienced the crisis as creditor states facing calls to support other Eurozone members (Frieden and Walter 2017; Laffan 2017). Given that the crisis had a more tangible negative impact on the crisis states, and in the light of the perception that the crisis response was largely imposed by the creditors, we hypothesise that the EU’s legitimacy should face more serious challenges in the crisis states.

H3

Negative assessments of the EU’s legitimacy are more widespread in crisis states than in non-crisis states.

The considerations presented so far do not imply that de-legitimation is an automatic process. Political actors, through their discursive activities, seek to drive the de-legitimation sequence to the next stage, or to prevent this from happening. In the EU context, populist parties on the fringes of the political spectrum can be expected to be most interested in de-legitimating European integration (Halikiopoulou et al. 2012). It is also often assumed that journalists have a tendency to focus on negative developments; they might hence be particularly prone to engage in de-legitimating discourse (Galpin and Trenz 2017). By contrast, mainstream parties and politicians—especially those serving in executive office at the EU or member-state level—seek to de-politicise potentially controversial issues, and if this is not successful, will rally to defend the political system’s legitimacy (Statham and Trenz 2015). It is hence plausible to expect differences in EU-related legitimation discourse between various groups of speakers.

H4

Negative assessments of the EU’s legitimacy originate primarily from journalists, opposition politicians, and civil society speakers, while politicians acting in executive function at the EU and member-state level tend to defend the EU’s legitimacy.

Our fifth and final hypothesis examines whether different ways of framing discussions about the Eurozone crisis have an impact on the resulting legitimacy debates. As we discussed at the beginning of this article, the crisis has often been perceived as being particularly damaging for the EU’s legitimacy because it did not only entail economic problems, but also raised political concerns about the democratic and legal quality of EU rule (Mayer 2012, pp. 103–116; Crum 2013; White 2015). We assume that crisis-related arguments framed in such terms are particularly likely to result in negative assessments of the EU’s legitimacy.

H5

Negative discourse about the legitimacy of the EU is more widespread in arguments that reach beyond a purely economic assessment of European integration.

In the remainder of this article, we will rely on these five hypotheses to assess whether the EU’s legitimacy has become more precarious during the crisis, and which discourse constellations have tended to trigger de-legitimating political discourse.

Methodology: analysis of political claims in newspapers

The empirical analysis of legitimacy debates can, in principle, focus on different arenas of political discourse—including parliaments, internal forums within parties or associations, the news media, or the Internet. Each of these arenas influences legitimacy debates with their own selection and framing biases. This study concentrates on the news media and specifically on quality newspapers. Our methodological choice is informed by a number of considerations. First, the news media is a discursive arena in which the positions of both EU and member-state politicians are strongly represented, but which makes it possible to compare their interventions with those of professional observers (journalists and other commentators) and speakers from civil society. In other words, the focus on the media—rather than parliaments, for instance—allows us to identify what is unique about the interventions by different types of speakers. Second, examining the quality press ensures a relatively high density of reporting on European affairs, given that quality newspaper tends to be more Europeanised than tabloids (Kleinen-von Königslöw 2012). We also opted against tabloids to reduce the negativity bias in our material, assuming that if de-legitimation tendencies become evident in quality newspaper reporting, this constitutes a stronger indicator for legitimacy problems than if the same happens in tabloids. In the light of these considerations, it should be clear that the material we studied is elite-dominated; our results do not allow for inference on other political arenas, let alone on debates among ordinary citizens.

This study focuses on debates in four Eurozone member states: Germany, Austria, Spain, and Ireland. While we cannot assume that these states are representative for all Eurozone states, our case selection is intended to reflect discursive dynamics in both crisis and non-crisis states. Germany and Austria were two of the main creditor states which financed a large share of the bailouts and whose governments became advocates for the associated austerity conditions. The two countries differ in size and political influence, and also because explicitly Eurosceptic political positions have traditionally been represented more strongly in the Austrian than in the German political system (where they gained popularity more recently). Spain and Ireland, on the other hand, were hit hard by the crisis and were ultimately forced to accept bailout funds. These crisis states differ, once again, in size and in perceptions of European integration, with Ireland often assumed to be particularly EU-friendly. The newspapers included in the study are Süddeutsche Zeitung and Frankfurter Allgemeine Zeitung for Germany, Der Standard and Die Presse for Austria, El País and El Mundo for Spain, as well as Irish Times and Irish Independent for Ireland. This sample of newspapers represents one paper with a centre-left and one with a centre-right editorial line in each country.

Our analysis focuses on media debates five days prior and five days after the regular June and December meetings of the European Council between 2009 and 2014. In all of these meetings, coping with the financial crisis was a major agenda item. While media reporting is therefore likely to focus on policy-oriented concerns, the European Council’s high-profile character also makes its meetings a welcome occasion for newspapers to provide general reflection on the EU’s state of affairs. Our study does not encompass all European Council meetings that took place between 2009 and 2014, let alone all crisis-related media discourse. However, the selection of the regular June and December meetings allows us to establish a consistent sequence of time periods that makes it possible to assess change in the temporal dimension. As the timeline in Fig. 1 indicates, the period between 2010 and 2013 were the years in which the crisis was most intense. The first year and the last year in our study—2009 and 2014—provide insights into EU-related discourse in years in which the crisis was not as dominant. Incidentally, both years also featured an EP election.

Fig. 1
figure 1

Timeline of major events in the Eurozone crisis (X = time period analysed in this study)

Full size image

The analysis that follows draws on a larger study that examined all EU-related articles published during our time periods in one of our eight newspapers. We analyse these articles using a variation of political claims analysis, an approach first developed for the Europub project at the Social Science Research Centre Berlin (Koopmans and Statham 1999; Koopmans 2002). The basic unit of analysis in this approach is a political claim, defined as the public expression of an opinion related to institutions, processes, or results of collective decision-making. Claims represent explicit and purposive discursive action. They can take the form of political evaluations (support/criticism), political demands, or both.Footnote 2 Claims can originate from journalists or other authors of newspaper articles as well as from speakers cited in them. This study was interested only in claims related to the EU or European integration; claims about EU member states were treated as relevant if they concerned domestic policies linked to European integration (such as the government’s negotiation strategy for an EU summit) or if the claimant established an explicit link between domestic objects and the EU (for instance, by arguing that a domestic policy violates EU law). To identify articles with relevant claims, we first ran an automatic searchFootnote 3 and then manually screened all pre-selected articles. Following this, all claims contained in the selected articles were coded using the following main categories:

  • Claimant The author/speaker of the claim. The codebook distinguished between different types of claimant, most importantly journalist/guest author, EU/European politician (subdivided in EU executive actor and other EU actor), member-state politician (subdivided in member-state executive actor and other member-state actor), international actor (representing a third country or international organisation), and non-governmental speaker.

  • Object The aspect of the EU/European integration that is discussed in the claim. Since our study is interested in the distinction between polity- and policy-related contestation, this category was defined in a way that encompasses both aspects of EU-related discourse. Claims were coded on whether they referred to the EU polity/EU institutions, EU policies, or member-state policies related to the EU.

  • Evaluation An explicit negative or positive assessment of the object. The analysis presented here focuses only on evaluations of an object’s status quo (not evaluations of proposals for how the object could be changed). Only negative or positive evaluation was coded; there was no ‘neutral’ category.

  • Demand A call for (or against) specific political activities by the EU or its member states relating to the object. As noted above, a claim in our conception can be a pure evaluation (with no demand), a pure demand (with no evaluation), or a combination of evaluation and demand.

  • Justification The kinds of reasons (if any) presented for an evaluation or demand. In this respect, we distinguished between economic arguments, other pragmatic/outcome-oriented arguments such as national interests or policy-specific objectives, and moral or ethical arguments such as democracy, legality, or identity (for this distinction, see Habermas 1993).

In total, we identified 6069 claims (1954 from Germany, 1367 from Austria, 1929 from Spain, and 819 from Ireland), drawn from 3801 relevant newspaper articles, and coded them based on the above categories.Footnote 4 Table 2 provides examples.

Table 2 Examples of media statements and coding
Full size table

The coding generates measures to operationalise the categories developed in the theoretical section of this article and to test our five hypotheses. The overall number of claims, especially when put in relation to the total number of articles published in any given period, measures the salience of European integration and hence gives insights into levels of politicisation. However, not all of claims have a legitimacy component. In our operationalisation, legitimacy claims are claims that (a) focus on the object category EU polity/institutions, rather than policy-oriented objects and (b) are evaluative in character, rather than being pure demands. In Table 2, only the third example qualifies as a legitimacy claim.

Based on this conception, we can track the development of legitimacy debates over time (H1) by examining the absolute and relative frequency of legitimacy claims; we can also identify which share of legitimacy claims is positive (legitimating) and negative (de-legitimating) (H2). In addition to this descriptive analysis, we can calculate a regression analysis that gives insights into the discourse constellations that influence legitimacy claims, by exploring how they are shaped by the countries from which the originate (H3), the type of speaker/claimant who presents them (H4), and the kinds of justifications that are used to support them (H5). Differences between paper types (centre-right or centre-left) and article types (news reports or opinion articles) are included in this analysis as control variables.

Developments over time: from a policy debate to a legitimacy debate?

In a first step of our inquiry, we were interested in legitimacy trajectories over time. Did the frequency of legitimacy claims increase over the course of the Eurozone crisis? Did legitimacy claims turn more negative? Answering these questions allows us to assess whether the hypothesised shift from a policy debate to a legitimacy debate about the EU did indeed occur.

Table 3 provides an overview of the different kinds of claims distinguished in our analysis. It shows that EU-related claims make up a relatively small share of reporting; on average, we identified less than four claims for every 100 newspaper articles printed in the periods we examined. Approximately half of these claims, with some variation between countries and over time, have an evaluative component; the others express demands without an explicit evaluation. Most of these evaluations do not concern the EU polity or EU institutions, but rather address European or domestic policy issues. Only approximately one in ten claims is a legitimacy claim. Of these legitimacy claims, slightly more than half are de-legitimating.

Table 3 Types of claims over time
Full size table

Our first two hypotheses (H1 and H2) focus on change over time. To facilitate the analysis of temporal patterns in each country, Figs. 2, 3, and 4 display developments based on a standardised scale, with 2009 as the base year. Figure 2 maps the overall number of claims, calculated in relation to the number of articles published in the selected time periods. This figure provides a measure of the salience of European integration in political discourse, a core dimension of politicisation. It reveals similar trajectories in all four countries: the share of EU-related claims rises until 2011, when it reaches a peak, and then declines to levels close to, or even slightly below, the starting values. This finding is in line with previous research that has found an intensification of EU-related media debates during the crisis (Kriesi and Grande 2016; Hutter and Kriesi 2019). However, it also shows that the increase in attention devoted to the EU is a temporary phenomenon that passes once the crisis subsides.

Fig. 2
figure 2

EU-related claims over time (share of articles published, 2009 = 100)

Full size image
Fig. 3
figure 3

Legitimacy claims over time (share of all claims, 2009 = 100)

Full size image
Fig. 4
figure 4

De-legitimating claims over time (share of legitimacy claims, 2009 = 100)

Full size image

As we have argued in the theoretical section, such politicisation trajectories do not need to be reflected in legitimacy debates. Our first hypothesis (H1) rather expects an increase in legitimacy debates over the course of the crisis, as its impacts accumulate and inflict more and more damage on European integration. As Fig. 3 reveals, only debates in Germany are characterised by a consistent trend towards more intensive legitimacy debates. In the other states, there is no noticeable increase in legitimacy debates until 2014, the last year of our study. Nevertheless, it is noteworthy that, in all countries except Ireland, legitimacy claims make up a higher share of EU-related newspaper discourse at the end of our time period than at the beginning. The exceptionality of Ireland is likely due to debates about the country’s ratification of the Lisbon Treaty in the run-up to the second referendum on the issue in October 2009. These debates kept alive legitimacy issues linked to the treaty change (Vetters et al. 2009; Hurrelmann et al. 2013) which in the other member states had largely subsided by 2009. With this qualification, we can conclude that H1 finds general support in our material.

Similar discourse patterns emerge if we examine the share of de-legitimating statements. Our second hypothesis (H2) suggests that legitimacy debates should turn more negative over time. Figure 4 shows that there is indeed a slight increase in de-legitimation. In all countries, de-legitimating claims make up a higher share of legitimacy claims at the end of our time period than at the beginning. However, this development is driven primarily by discourse in 2014. In earlier years, we observe some fluctuations, but no clear trends. These findings do not disprove H2, but they suggest that discourse dynamics in 2014 require further analysis.

In conclusion, our analysis of temporal patterns reveals that there was indeed an increase in EU-related legitimacy debates, and a growing share of negative legitimacy evaluations, in the crisis years. However, these developments were not particularly pronounced and occurred rather late in the six-year period examined here, especially in 2014.

Discourse constellations: which factors make de-legitimation more likely?

Which factors account for these developments? In addition to providing insights into trajectories of legitimacy debates, our coding is designed to identify discourse constellations that lead to challenges to the EU’s legitimacy. Hypotheses H3–H5 highlight three explanatory factors: differences between the countries examined in this study (H3), types of claimants (H4), and justifications used to frame evaluations of the EU polity and its institutions (H5).

Table 4 presents descriptive data on these variables. With respect to countries, it shows that de-legitimating claims make up a larger share of EU-related discourse in Austria and Spain than in Germany and Ireland. Regarding claimants, it shows that de-legitimation occurs more frequently in claims by journalists and civil society speakers than in claims by EU executive actors, member-state executive actors, and international representatives (that is, leaders of non-EU states or international organisations). Claims by EU and member-state politicians who do not serve in executive function fall in between both extremes. Regarding justification, we see that claims that are framed with reference to moral or ethical standards such as democracy, legality, or collective identity are more frequently de-legitimating than claims framed with reference to economic considerations. Claims that refer to other pragmatic justification standards, such as the protection of national interests, fall in between both extremes.

Table 4 Types of claims by country, claimant, and justification
Full size table

The combined effect of these three factors on the likelihood of de-legitimation can be estimated using a binary logistic regression analysis. This procedure models the influence of a set of predictors (independent variables) on a binary outcome (dependent variable). In our regression model, the outcome we are interested in is the likelihood of a claim being de-legitimating. Predictor variables are the three factors referenced in H3–H5 (country, claimant, and justification). The year of the claim, paper type (centre-left or centre-right), and article type (news report or op-ed) were included as control variables. Binary logistic regression analysis expresses the influence of independent and control variables on the dependent variable in the form of an odds ratio, which indicates how a change in the value of the predictors affects the odds of a particular outcome occurring—in this case, a de-legitimating claim being made. An odds ratio higher than one indicates that as the predictor changes, the odds increase. Conversely, an odds ratio lower than one points to decreasing odds of de-legitimation.

Table 5 presents our regression results.Footnote 5 A look at the variables that are statistically significant allows us to test our explanatory hypotheses:

  • With respect to country (H3), we find that de-legitimating claims are comparatively more likely in Austrian and Spanish newspapers. When compared to our reference category, Germany, and assuming that all other variables are held constant, the odds of a de-legitimating claim increase by 68.5% if a claim stems from an Austrian newspaper and by 76.6% if a claim stems from a Spanish newspaper. The odds ratio for Ireland is not statistically significant. This finding contradicts H3, which suggested that the states most directly affected by the crisis—Spain and Ireland—should have the most negative discourse. What it suggests is that economic effects of the crisis are mediated by other factors, including perhaps long-standing attitudes towards European integration (which have been much more EU-friendly in Ireland than in Austria).

  • With respect to claimants (H4), we see that de-legitimation is comparatively less likely if a claim is made by politicians serving in executive function at the EU or member-state level, or in a third country or international organisation. Compared to our reference category, claims by journalists, and if all other variables are held constant, the odds of de-legitimation decrease by 76.6% if a claim is made by an EU executive actor, by 38.8% if a claim is made by a member-state executive actor, and by 64.1% if a claim is made by a third country or international actor. This is in line with H4, which expressed the expectation that politicians representing EU institutions and national governments would be particularly prone to defending the EU’s legitimacy.

  • For justification (H5), the regression analysis shows that de-legitimation is comparatively less likely in claims that are framed in economic terms and more likely in claims that are framed in moral or ethical terms. Compared to claims in which no justification is provided, and if all other variables are held constant, the odds of de-legitimation decrease by 72.8% if a claim uses an economic justification and increase by 68.8% if a claim uses a moral or ethical justification. As hypothesised in H5, this finding suggests that negative assessments of the EU’s legitimacy are frequently driven by concerns about the EU’s legal and democratic quality, while purely economic arguments are less likely to trigger de-legitimating discourse.

Table 5 also shows that two of our control variables, article type and year, are statistically significant. Regarding article type, it is unsurprising that claims made in opinion pieces—which may be authored by journalists, but also by politicians, representatives of non-governmental organisations, or academic experts—are more likely to be de-legitimating than claims contained in news reports. Regarding year, our regression analysis once again indicates the exceptional character of discourse in 2014. If all other variables are held constant, the odds of a claim from 2014 being de-legitimating are 127.5% higher than in the reference year 2009. Similar tendencies can be found in no other year. In other words, even if we control for all other variables that have a significant impact on the likelihood of de-legitimation, developments in 2014 still stand out. This indicates that legitimacy discourse in 2014 deserves to be examined in a more fine-grained analysis.

Table 5 Factors that influence the odds of de-legitimating claim (regression analysis)
Full size table

What was special about 2014? Legitimation and leadership selection

What accounts for the particularly high levels of de-legitimating discourse in 2014, evident both in our timelines and in our regression analysis? We approach this question through a qualitative analysis, designed to identify aspects of media reporting that are not covered by our hypotheses. This last step of our analysis is based on the 130 legitimacy claims identified for 2014, particularly the 85 de-legitimating claims. It is noteworthy that most of these originated from debates about two developments: the May 2014 EP election and the formation of a new European Commission that followed. The EP election resulted in unprecedented successes of Eurosceptic parties in some member states, especially France and the UK. To many observers, this indicated that the Eurozone crisis had indeed led to a pervasive erosion of citizen trust in European integration. While this development had been predicted, inter alia by Commission President Barroso in the statement cited at the beginning of this article, it had not featured prominently in newspaper discourse in our four countries until 2014. After the election, newspaper commentators raised the question of whether the EU’s increasing disconnect from the citizens, as evidenced by the election result, constituted a fatal flaw of the integration project. In one typical op-ed, a commentator in Austria’s Die Presse put it as follows:

For a long time, the EU lived off a positive prejudice, supported by socio-economic progress, the freedom gained by tearing down borders, and so on. In the course of the crises, this sentiment has turned on its head: Today, a negative prejudice dominates. All problems are blamed on Brussels, even though the member states caused many weaknesses of the EU with their half-hearted approach to integration, or they are themselves the reason for the crises. […] The European election 2014 has made the threat to the EU impossible to deny. The EU project can only be realised with the support of the citizens, but it can only achieve this support if it has positive effects. This in turn requires reforms and an increase in integration, both of which are impossible without democratic support. A vicious circle, in which the EU project may be destroyed (Die Presse, June 25, 2014; translated by the authors).

The second development that triggered a large share of de-legitimating claims in 2014 was the formation of a new European Commission following the election. 2014 was the first year in which the so-called Spitzenkandidaten process was used, meaning that the main party families put forward candidates for the office of Commission President. The candidate of the strongest EP group, Jean-Claude Juncker of the European People’s Party, was subsequently nominated by the European Council as Commission President. However, the institutional logic of this process was opposed by Hungary’s Prime Minister Viktor Orbán and, even more vocally, British Prime Minister David Cameron. Cameron’s complaints about Juncker’s nomination were not only reported widely in all four countries, but also resulted in broader debates about the EU’s institutional processes for selecting its leaders, and whether they are adequate to deal with ongoing challenges (including the Eurozone crisis). The following three excerpts illustrate some of the de-legitimating arguments brought forward in this context:

British officials stressed yesterday that the election formula for the president of the common executive, devised by the European Parliament and used for the first time in the elections of May 25, with its pre-designated candidates for the Commission leadership, means that ‘very good’ candidates are ‘automatically’ discarded. (El País, June 26, 2014; translated by the authors)

A Commission that owes its position to the parliamentary parties will not be able to make use of its powers of control and leadership. The Council will watch it wearily; it will not forgive the European Parliament and the Commission President their assertiveness. All the while, the EU is in a deep crisis. But this crisis has not led to a push for institutional reform. All that has grown is a desire to engage in theatrics. (Süddeutsche Zeitung, June 26, 2014; translated by the authors)


Since the beginning of the crisis, the European project has increasingly fallen into the hands of ideologues supporting redistribution and state intervention. The biggest problem […] are the EU institutions themselves. Cameron is correct when he calls Brussels a ‘parallel universe’. […] The Frankfurter Allgemeine Zeitung recently spoke in a commentary of a ‘spaceship Brussels’: ‘The self-centredness of the Brussels institutions is the real democratic deficit of the EU.’ (Der Standard, June 28, 2014; translated by the authors)

The legitimacy discourse triggered by processes of leadership selection in the EU—the EP election and the nomination of the new Commission President—underlines that legitimacy is strongly tied, conceptually but also in the way it is empirically constructed, to questions of political power. Our qualitative analysis suggests that de-legitimation increased in 2014 not primarily because the effects of the Eurozone crisis were more strongly felt than in previous years, but because this year provided an institutional constellation in which crisis-related concerns could be discursively linked to the question of who should exercise power in the EU and how power-holders should be selected. What our original hypotheses miss, in other words, are variables that focus on institutional procedures of leadership selection, around which legitimacy debates crystallise. This does not mean that the crisis itself was irrelevant for legitimacy discourse. The year 2009, for instance, also featured an EP election and the formation of a new Commission, but de-legitimation was much lower. What we have seen, however, is that institutional factors need to be taken seriously as catalysts that help a policy crisis develop a legitimacy dimension.

Conclusion

There can be little doubt that the Eurozone crisis was an important juncture in the history of European integration. A number of studies have shown that, in addition to having far-reaching effects on EMU policy, the crisis gave a boost to the politicisation of EU governance, by triggering EU-related debates in a variety of discursive settings, including the news media (Kriesi and Grande 2016; Leupold 2016; Hutter and Kriesi 2019). The objective of this article was to explore whether this crisis-induced politicisation remained restricted to EU policy or whether media debates were also characterised by a trend towards increasingly frequent—and increasingly critical—discussions about the legitimacy of the EU polity. In the light of political and scholarly debates about an alleged ‘legitimacy crisis’ of the EU (Longo and Murray 2015; Schweiger 2017), we hypothesised that such de-legitimation tendencies should be observable in media discourse. Yet, our results show that this was the case only to a limited extent. There was no pronounced shift from policy debates to legitimacy debates; we also did not find an obvious correlation between a Eurozone state’s crisis experience and the likelihood of de-legitimation. Indeed, it was not until the EP election of 2014, when the strong showing of Eurosceptic parties made citizen discontent with the EU impossible to ignore, that the frequency of de-legitimation statements in the newspapers studied here began to increase.

In interpreting these findings, it is important to acknowledge the limitations of our study. First, we examined only four Eurozone states; we cannot rule out that media debates in others—Greece, for instance—where characterised by much more explicit de-legitimation. Second, as we have stressed in this article, our research on legitimacy debates in quality newspapers cannot be generalised to other discursive arenas, let alone to the population at large. It does, however, provide useful insights into the conditions under which media reporting picks up on, and potentially serves as a catalyst for, challenges to the EU’s legitimacy. As we have seen, this happens most frequently in claims that are made by non-politicians (such as journalists and civil society), that frame assessments of the EU with reference to moral or ethical arguments (such as democracy, legality, or collective identity), and that establish a discursive connection to processes of leadership selection (such as elections or nomination procedures for top offices). All of this implies that as EU politics becomes less technocratic and more competitive in character, legitimacy challenges in media discourse are likely to increase.

In the Eurozone crisis, this happened only to a limited extent. Newspaper reporting remained dominated by politicians; economic or other pragmatic arguments were the most popular; and opportunities to link the crisis to questions of political leadership did not develop, at the EU level, until 2014. Media discourse thus remained out of step with perceptions of the crisis as a legitimacy challenge to the EU. This likely facilitated the ability of EU elites to develop technocratic responses to the crisis (Schimmelfennig 2014; Börzel and Risse 2018). The fact that no legitimacy crisis developed in the media arena also implies, however, that any legitimacy concerns that the crisis may have raised in the eyes of the public remained unaddressed.