Public opinion and the crisis: the dynamics of support for the euro

ABSTRACT Further integration in the European Union (EU) increasingly depends on public legitimacy. The global financial crisis and the subsequent euro area crisis have amplified both the salience and the redistributive consequences of decisions taken in Brussels, raising the question of how this has influenced public support for European integration. In this contribution, we examine how public opinion has responded to the crisis, focusing on support for monetary integration. Interestingly, our results show that support for the euro has remained high within the euro area; however, attitudes are increasingly driven by utilitarian considerations, whereas identity concerns have become less important. While the crisis has been seen to deepen divisions within Europe, our findings suggest that it has also encouraged citizens in the euro area to form opinions on the euro on the basis of a cost–benefit analysis of European economic governance, rather than relying primarily on national attachments.


INTRODUCTION
The global economic crisis and the ensuing euro area crisis have highlighted the salience and the redistributive consequences of monetary integration. As a consequence, citizens have become increasingly aware of the interdependence of European economies and the importance of European Union (EU) institutions in determining the future of individual national economies. Institutional responses to the crisis have also brought about reforms to deepen integration in the EU and resulted in calls for greater democracy and enhanced legitimacy. The euro area crisis thus represents the clearest example to date of Europe-wide politicization of the integration issue, understood in Schmitter's sense as a 'widening of the audience or clientele interested and active in integration ' (1969: 166). This can thus be seen as a critical step in the ongoing process leading away from the 'permissive consensus' of the early period of integration, where insulated leaders could make decisions without public consultation, towards a 'constraining dissensus' where public opinion is both more critical and more decisive (Hooghe and Marks 2009). With a proliferation of referendums on EU matters, politicization of EU issues in national elections and increased powers of the European Parliament, future reforms to deepen European integration hinge more than ever upon public support. This raises the question of how greater public awareness of European integration and its economic consequences changes the nature and role of public opinion in the integration process. In this contribution, we address a key aspect of this question by examining how the dynamics of support for the EU have been shaped by the euro crisis. Specifically, we analyse how attitudes towards monetary integration have been influenced by the global financial crisis and the euro area crisis since 2008. Our findings demonstrate that, contrary to what one might expect, public support for economic integration has remained stable within the euro area, while it has declined in EU member states outside the euro area (see also Hobolt and Leblond [2014]; Roth et al. [2011]). However, the factors that shape euro support in the euro area have shifted from identity-based concerns before the crisis to more utilitarian considerations during the crisis. These results suggest that as citizens receive more information about the economic consequences of monetary integration, people's attitudes towards the euro are more likely to be based on a systematic evaluation of the costs and benefits of the integration process. Such a utility-based approach to integration implies that citizens assign greater weight to individual and societal economic benefits from integration when forming opinions about monetary integration and are less likely to oppose the euro because of their attachment to the nation. Hence, while support for the euro may appear stable, the continued legitimacy of EU economic governance could depend on whether the EU is seen to deliver effective solutions to economic problems.

PUBLIC OPINION AND EUROPEAN INTEGRATION
Grand theories of integration have largely neglected the role of public opinion, not least because the general public was considered to be of minimal importance for the process of integration. However, both major integration theories, liberal intergovernmentalism and neofunctionalism, allow for ways in which public opinion could potentially become relevant. Liberal intergovernmentalism theorizes the importance of domestic interests in shaping governments' preferences as a first 'level' of the two-level game of European integration. This could potentially provide a role for public opinion, assuming that EU matters are sufficiently salient to influence the electoral incentives of national politicians (Börzel and Risse 2009;Moravcsik 1998). Yet, the primary focus of liberal intergovernmentalism is domestically organized economic interests shaping national governments' positions, rather than the role of national electorates (Moravcsik 1998). Similarly, according to neofunctionalism, politicization of EU issues could render public opinion important; yet other drivers of integration, notably economic interests organized in transnational associations and the preferences of member states, are considered far more significant (Haas 1958;Hooghe and Marks 2006;Schmitter 1969). Hence, while grand theories have acknowledged that public opinion can potentially play a role if the integration process became more salient, they have only paid limited attention to the mechanisms through which it should become important, and they have not engaged with the question of how public opinion is shaped by the process of integration itself.
The increasing politicization of the EU since the 1990s, however, has led scholars to pay more attention to the impact of public opinion on integration. Notably, Liesbet Hooghe and Gary Marks (2009) have formulated an ambitious 'postfunctionalist theory', which incorporates the role of public opinion into a framework for understanding regional integration. They argue that more democratic control over EU decision-making and increased politicization of the EU issue in domestic party politics have brought the public into EU decisionmaking. Their focus is on the conditions under which an EU issue becomes politicized in the domestic context and the move from the 'distributional logic' to the 'identity logic' of integration. They argue that as the scope and depth of European integration has intensified and the tension over threats to national identity have become more salient, not least due to political entrepreneurs mobilizing the issue, identity also becomes the key factor shaping attitudes towards integration (Hooghe and Marks 2009).
This argument relates to the ongoing debate between two alternative perspectives on public support for European integration: a utilitarian and an identitybased approach. From a utilitarian perspective, generic support for European integration is determined by a rational cost -benefit analysis: those who benefit economically from European integration (particularly trade liberalization) are supportive, whereas those who stand to lose are more hostile (Gabel 1998;Gabel and Palmer 1995;McLaren 2006). Support for monetary integration has also been explained in utilitarian terms, arguing that individuals with high involvement in international trade should favour the euro more than individuals employed in the non-tradable sector (Banducci et al. 2009;Gabel 2001;Gabel and Hix 2005). Studies of support for the euro have also found that sociotropic economic concerns play a role: citizens in countries that benefit economically, or are perceived to benefit economically, from membership of the EU are more supportive of the euro (Banducci et al. 2003(Banducci et al. , 2009Hobolt and Leblond 2014;Kaltenthaler and Anderson 2001). An alternative explanation for the variation in support for European integration, and the euro more specifically, focuses less on economic self-interest and more on the threat that European integration can pose to national identity and a country's symbols and values (Carey 2002;Hooghe and Marks 2004;McLaren 2006). Several studies have shown that attachment to the nation, and particularly exclusive national identity, is a powerful predictor of negative attitudes towards European integration (Hooghe and Marks 2004;McLaren 2006). In the context of the referendums on joining the euro in Denmark and Sweden, Jupille and Leblang (2007) found that 'identity concerns' played a greater role than 'pocketbook calculations'. Generally, citizens who thought that the EU undermined national sovereignty and democracy were more likely to vote against the euro's adoption (see also Hobolt and Leblond [2009]). In their recent study of support for European economic integration during the crisis, Kuhn and Stoeckel (2014) also show that both utilitarian considerations and national identity matter.
Hence, existing work has clearly demonstrated that economic calculations and identity concerns play an important role in explaining variation in support for the EU, and monetary integration more specifically. However, less work has examined how the balance between these concerns evolves once the integration process and its economic consequences become more salient. 1 The visibility of European economic integration undeniably increased during the euro crisis, as Kriesi and Grande (2014) have shown in their content analysis of the public debate in six Western European countries. They conclude that 'the debate [on the euro crisis] has been exceptionally salient and has contributed to the increased visibility of Europe in the politics of the European nation-states' (Kriesi and Grande 2014: 24). While previous bursts of interest in the EU have tended to concern constitutional matters (e.g., treaty changes) and be concentrated in specific countries (e.g., where referendums were held), the euro crisis is unique in that it has made the issue of European integration salient across Europe and that it has highlighted decisions at the European level that have very obvious redistributive consequences between and within countries (Cramme and Hobolt 2014). Since the crisis has highlighted the tangible economic consequences, we expect that it is unlikely to evoke an 'identity logic', but that it is rather likely to prompt citizens to employ a 'utility logic'. Below, we elaborate on this idea and develop our expectations concerning the public response to the crisis.

THE PUBLIC RESPONSE TO THE CRISIS
Citizens' response to the 'Great Recession' needs to be understood in the context of the increasing salience and public awareness of the consequences of monetary integration. The heightened public salience of integration can be attributed to a mixture of factors: the visibility of the euro crisis and the EU's attempt to deal with it; changes in individual circumstances linked to the crisis (e.g., unemployment); as well as deliberate attempts by certain political élites to mobilize the issue and thus expand issue-specific conflict beyond the narrow circle of political actors (Cramme and Hobolt 2014;De Vries and Hobolt 2012;Hooghe and Marks 2009). Moreover, in response to the crisis the EU made a number of interventions that can be seen to increase integration in the EU, including measures that were targeted at individual debtor states and more formal institutional reforms to deepen economic integration in the EU (see Ioannou et al. [2015]).
There is evidence that the crisis raised the public awareness through all of these mechanisms: initial awareness for the interdependencies monetary integration had created was already present at the outset of the crisis and this awareness became more acute after the reforms of European economic governance, when the issue was also politicized by political entrepreneurs. The question of the desirability and future of economic integration in the Union became commonplace in the national public spheres of euro area countries and a key feature of national party competition and domestic election campaigns (Cramme and Hobolt 2014; Kriesi and Grande 2014). As the crisis evolved from a financial crisis into a sovereign debt crisis and a euro area crisis, citizens also started to assign responsibility to the EU for its actions and inactions. These shifts in public awareness of the EU's impact on the national economies were particularly pronounced inside the euro area (Hobolt and Tilley 2014).
Our expectation is that as monetary integration became an issue of general public awareness and information on the issue was disseminated, the nature of opinion formation also shifted. We know from the literature in political psychology and behavioural economics that greater salience allows people to engage in more systematic and analytical evaluations, and causes them to rely less on predispositions and heuristics. Negative information in particular has been shown to encourage individuals to engage in systematic evidence-based opinion formation (Kahneman and Tversky 1979). Systematic processing means that individuals take heed of the decision-relevant information that is currently available, and based on this information they carefully piece together a decision. This can be contrasted with heuristic processing which requires much less effort by individuals, as new information is processed in accordance with these standing decisions, including affective attachments, such as identities or partisanship, and not taking into account new information (Chaiken 1980).
In the context of the EU, it has been shown that issue-based proximity voting, rather than protest voting, is more common in European elections and referendums when the EU issue receives more media coverage and is more salient in the political debate (de Vries 2007;Hobolt 2009;Tillman 2004). Equally, since the Great Recession made European integration, and particularly monetary integration, a more salient issue in the national public debates, we expect citizens to become more concerned with economic pros and cons of integration. Consequently, our expectation is that utilitarian considerations concerning the economic benefits from economic integration and the capacity of institutions to deliver these benefits have come to matter more and identity to matter less as the issue of economic governance has become more salient. We expect this to be the case primarily inside the euro area, where there was a fear not only that individual economies might default, but that the entire system would collapse, thus creating great uncertainty and risk for euro area members regardless of whether they were debtor or creditor states. In contrast, we do not expect change outside the euro area, where the economic implications of a potential euro area meltdown were more diffuse and received less public attention. We focus on support for the type of integration most clearly linked to the crisis, namely support for monetary integration and the euro. This leads us to the following hypotheses: H1: Economic benefits of integration have become more important to citizen support for the euro inside the euro area after the onset of the economic crisis. Outside the euro area the importance of economic benefits has remained stable.
H2: Institutional capacity to deliver benefits from economic integration has become more important to citizen support for the euro inside the euro area after the EU commenced its institutional reform agenda in response to the crisis. Outside the euro area the importance of institutional capacity has remained stable.
H3: National identity has become less important to citizen support for the euro inside the euro area during the crisis compared to before the crisis. Outside the euro area the importance of national identity has remained stable.
To test these propositions we start by looking at aggregate-level trends in support for the euro, and thereafter we present the results from multilevel models of pre-crisis and crisis surveys of individual support for the euro.

MEASURING THE PUBLIC RESPONSE TO THE CRISIS
It is not straightforward to put a date on the beginning of the European economic and financial crisis. While the onset of the financial crisis is often dated back to the collapse of Lehman Brothers in September 2008, survey data show that Europeans were aware of the looming economic crisis as early as the beginning of 2008: the proportion of citizens of Europe thinking that the 'economy will get worse' increased by a remarkable 20 percentage points between the autumn 2007 and the spring 2008 Eurobarometer survey waves (see Eurobarometer surveys). 2 The so-called 'euro area crisis' only emerged a little later. At the heart of this evolution was the sovereign debt crisis that surfaced in 2009 with downgrading of government debt in many European states, particularly in the so-called 'GIIPS' countries (Greece, Ireland, Italy, Portugal and Spain). Concerns intensified in early 2010 and thereafter, leading the EU to implement a series of financial support measures such as the European Financial Stability Facility (EFSF) and specifically 'bailout funds' to countries facing a severe sovereign debt crisis. These euro rescue measures targeted at helping countries in a severe sovereign debt crisis were accompanied by more formal institutional reforms of the governance of the Economic and Monetary Union (EMU), including the decision by the ECB to undertake outright monetary transactions and the establishment of the European Stability Mechanism to safeguard and provide instant access to financial assistance programmes (see Ioannou et al. [2015]). Second, a series of new legal instruments (the socalled 'six-pack', the 'two-pack', the Macroeconomic Imbalances Procedure), new decision-making procedures (the European Semester) and a new intergovernmental treaty, the Fiscal Compact, were aimed at more tightly constraining national fiscal policy-making (see Ioannou et al. [2015]). These ongoing attempts to rescue countries on the brink of bankruptcy, and avoid the collapse of the euro area, as well as the more formal institutional changes, were extensively covered in the national media across Europe.
To examine the public response to the crisis and the reforms, we rely on data from the Eurobarometer (EB), which are conducted twice a year on behalf of the European Commission, surveying citizens in each of the member states with respect to their opinions on European matters. According to our theoretical argument, citizens would have responded to the increasing salience of European economic governance by adjusting their opinions, but this response would not have been identical inside and outside the euro area, since the cost -benefit calculus for insiders and outsiders differs significantly. To explore these claims at the aggregate level, we measure public attitudes over time, both before and during the sovereign debt crisis in Europe, distinguishing between insiders and outsiders. We look specifically at support for the common currency, and we expect different responses inside and outside the euro area, since the utility calculus differs significantly depending on membership status: for countries that are already members, an exit or potential collapse of the euro area could have potentially disastrous and highly uncertain economic consequences; whereas for countries not yet members, the risk-adverse response may be to stay out of a currency area in turmoil. Public euro support is captured by way of the most commonly used to measure support for monetary integration, namely the proportion of citizens who favour the EMU with a single currency. Figure 1 shows, in line with our general argument, that support for the euro is lower outside the euro area than inside. We can also see that, whereas there has been a decline in support among euro outsiders since 2010, support inside the euro area has remained fairly stable (see also Clements et al. [2014]; Roth et al. [2011]). This suggests that the choice facing insiders and outsiders is very different, and whereas insiders have remained supportive, outsiders have become less certain about the benefits of joining a monetary union in crisis (Hobolt and Leblond 2014).
Moreover, despite the severity of the crisis, citizens across Europe continued to regard the EU as relatively more effective at dealing with the crisis than other institutions, such as national governments. We can demonstrate this empirically by using the Eurobarometer question: 'Which of the following is best able to take effective actions against the effects of the financial and economic crisis?' The options include the national government, the European Union, the United States, the G20, the International Monetary Fund (IMF), other, and none. Figure 2 shows the proportion of citizens who selected the EU as the institution best able to take effective action. It is worth noting that, on average, EU citizens perceived the EU to be most effective at taking action against the crisis compared to all other options, even national governments. Figure 2 shows that Europeans became more likely to name the EU as the most effective institution as the financial crisis evolved into a sovereign debt crisis in late 2009 and early 2010, especially inside the euro area, where the proportion of people mentioning the EU increased by 10 percentage points between May 2009 and May 2010. This increase also coincides with highprofile EU interventions, such as the first Greek bailout and the establishment of the EFSF to safeguard financial stability in Europe in May 2010. In other words, as the crisis worsened in Europe and the EU started to intervene, more people thought that the EU is best placed to deal effectively with the consequences of the crisis. This remained stable throughout the crisis, however, despite the increased measures adopted by the EU in response to the crisis.
Overall, these descriptive data in Figures 1 and 2 suggest that the crisis has not led to reduced support for the euro inside the euro area, whereas this crisis has led to a sharp decline in support for the single currency outside the euro area. These results suggest that the utility calculus differed inside and outside the EA after the onset of the crisis, and that for citizens in the EMU a euro area collapse or exit was seen as the potentially most costly option, not least due to the widespread perception that the EU can deliver the most effective solution to the crisis. To examine more rigorously whether such a 'utility logic' was applied by citizens inside the euro area, we use individual-level data from before and during the crisis.

MODELLING INDIVIDUAL-LEVEL EURO SUPPORT
To assess the changing determinants of euro support, we estimate identical models of euro support using several waves of pre-crisis and crisis data. This allows us to closely track changes in effects over time. To do this, we use several Standard Eurobarometer waves for the years 2005 to 2013. 3 For all these waves we estimate the effects of all determinants of euro support that were asked in the respective wave. Unfortunately, we cannot estimate all predictors in one model, as they are rarely included in the same wave, and we thus estimate slightly different models (see Web Appendix for details [online supplemental material]). 4 Our dependent variable in all models is the respondent's (dichotomous) answer to the question whether he or she is for or against 'A European economic and monetary union with one single currency, the euro' (see Figure 1). This question has the convenient property that it captures support for the euro meaningfully both for citizens inside and outside the euro area, which any question on adopting or reforming the euro could not achieve. Turning to the operationalization of utilitarian concerns, we are interested in finding variables that capture the economic benefits of integration (H1) and the capacity of institutions to deliver those benefits (H2). To capture perceived country-level benefits from integration, we measure the respondent's perception of whether his country has benefited from integration: 'Taking everything into account, would you say that (OUR COUNTRY) has on balance benefited or not from being a member of the EU?' Using this operationalization of country-level benefits raises a question about the causal direction, since the respondent's view on the euro may influence his evaluations of how much benefit the EU provides, as well as vice versa. However, since our core argument is not about the direction of causality but about a stronger cognitive link between cost-benefit analysis and euro attitudes due to the crisis, this is less of a concern. Even if perceived benefits are partly endogenous to euro attitudes, a stronger association during the crisis can be interpreted as evidence of a change in the nature of support for the euro. Secondly, to operationalize the capacity of institutions to deliver benefits, we use the Eurobarometer question on the respondent's perception of which actor is most able to tackle the effects of the financial and economic crisis, constructing a dummy variable for those who consider the EU to be the most effective of all actors (see also Figure 2).
Our third expectation is that national identity will matter less during the crisis (H3). To capture identity, we use the following question: 'In the near future, do you see yourself as (1) [nationality] only, (2) [nationality] and European, (3) European and [nationality], or (4) European only?' This question on personal self-image has been viewed as the gold standard for measuring national identity by many scholars (e.g., Marks [2004, 2005]). In a comparison of different identity measures, Sinnott (2006) has found that 'identification ratings', such as this question, have the highest levels of predictive validity for attitudes on the EU. We use a simple four-level scale indicating different 'degrees of national identity' with 1 for 'European only' and 4 for ' [nationality] only'. To test the importance of the specific national context, we also include a dummy for EMU membership. 5 In each model we include interactions between EMU membership and the independent variables. We also include standard controls for education, age and gender that have been shown to influence EU support.
Our estimates are based on multilevel logistic regression with random intercepts at the country level. By using logistic regression we essentially conceive of support for monetary integration as an unbounded, latent variable Y * that is linearly related to our independent variables and linked to the probability of supporting the euro through the logit link function (Pr Y ( ) = L Y * ( )). All estimated models then take the following general form: where X is the independent variable of interest, EMU is a dummy for euro area countries, u j is the random intercept at the country level, and 1 i the individuallevel error term. Our primary interest lies in the changing effect of the independent variables Xs on latent support for monetary integration Y * ( ) inside versus outside the euro area and before versus during the crisis. Therefore, we calculate the conditional marginal effects dY * dX of each of the explanatory variables X and their 95 per cent confidence intervals depending on EMU membership. Figures 3A to 3C show how latent support for the euro (in terms of log odds) changes in response to a unit change in X inside versus outside the euro area and for each available wave. These plots are based on the full regression models (available in the Web Appendix [online supplemental material]) and the conditional marginal effects (EMU versus not EMU country) are calculated from the interaction between the EMU dummy and the independent variable of interest (see Brambor et al. 2006). To further illustrate the changing effects of utility and identity concerns inside the euro area, we also plot the predicted probabilities of supporting the euro for a hypothetical individual with different values on the key independent variables before and over the course of the crisis (Figure 4). Starting with our first hypothesis (H1), recall that we posited an increase in the association between benefits from EU membership and support for the euro inside the euro area. Figure 3A shows strong support for this hypothesis. While the effect of the variable fluctuates considerably outside the euro area Figure 3B Determinants of euro support before and during the crisis: marginal effects of EU as most capable actor Figure 3A Determinants of euro support before and during the crisis: marginal effects of perceived benefits from EU membership Figure 3C Determinants of euro support before and during the crisis: marginal effects of different degrees of national identity Figure 4 Difference in predicted probabilities of supporting the euro for individuals with different values on the independent variable (euro area) Note: These figures show differences in predicted probabilities of supporting the euro for two hypothetical individuals (female, 47 years old, modal occupation and education level in a euro area country) who differ in terms of the variables of interest: perceived benefits from membership (benefits or no benefits); perceived EU effectiveness (EU as the most capable/EU not as the most capable); and identity ('national only' or 'national and European'). without a clear trend, the association has become much stronger within EMU over time. In particular, we witness a large shift in effects between the 2007 and the 2008 waves of the survey (pre-and post-crisis waves) that stabilizes and becomes even more pronounced in the subsequent years. This suggests the link between support for the euro and perceived national benefits from integration was changing from the very onset of the global financial crisis in 2008. This supports our conjecture that the crisis primed utilitarian concerns in citizens' reasoning about the euro. Indeed, the change is more immediate than one might have expected -it already occurred after the first signs of a crisis at the horizon. It coincides with a considerable worsening of citizens' economic expectations: whereas in autumn 2007 only about 26 per cent of EU citizens expected a worsening national economy in the next 12 months, this figure almost doubled to 46 per cent in just half a year. In accordance with our expectations, citizens in the euro area reoriented their stance on the euro in line with their assessment of whether integration provides benefits to their country. This is also reflected in predicted probabilities of supporting the euro: the difference between an individual who perceives benefits from integration and someone who does not was about 34-36 percentage points before the crisis but around 40 during the crisis (Figure 4). In other words, this factor clearly became more important in predicting support.
We find equally supportive evidence for our second hypothesis. Perceptions of the EU as the most effective actor to tackle the crisis became an increasingly important determinant of support for the euro as the crisis proceeded inside the euro area ( Figure 3B). Regrettably, we can only observe the change from spring 2009 onwards, when the question was asked for the first time. However, the subsequent waves clearly demonstrate that concerns about the EU's institutional capacity to deliver effective solutions become more important for euro support, in particular between 2009 and 2012, when the EU adopted a number of reform measures in response to the crisis. This change is statistically significant and it is also substantial in terms of predicted probabilities. Figure 4 demonstrates that inside the euro area those who viewed the EU as most effective actor had a 9 percentage points higher probability of supporting the euro in 2009 compared to those who did not view the EU as the best able to take effective actions, and notably this difference increased to over 15 percentage points in 2012 and 2013. In contrast, we do not see any change in the importance of institutional effectiveness outside the euro area, in line with our theoretical expectation. Clearly, citizens outside EMU had less reason to respond to the euro area crisis by engaging in a cost -benefit analysis of economic integration as they were less directly affected by the fate of the euro area reforms.
Lastly, we turn to changes in the effect of national identity. Figure 3C plots the effect of degrees of national identity on a four-point scale between 2005 and 2013. The findings strongly support our third hypothesis. The effect of the variable has significantly diminished within the euro area over time. Notably, identity plays a much less important role for supporting the euro in 2010 compared to 2005, and the effect of identity remains at this lower level in the following years. As with benefits from integration, this again corroborates the finding that the cognitive shift we observe already took place during the early phase of the crisis. However, as we have fewer data points for identity, we cannot ascertain when exactly the shift occurred. We can, however, conclude that citizens clearly relied less on identity heuristics when forming their views on the euro as the crisis unfolded. Figure 4 shows that, before the crisis, individuals in the euro area conceiving of themselves as 'national only' had about 21 percentage points lower probability of supporting the euro than those with 'national and European' identity. In contrast, this difference due to identity diminished to 14-17 percentage points during the crisis. Hence, the role of identity in predicting support for the euro diminished during the crisis. Outside the euro area, however, the effect of identity remained largely stable.
Overall, we find that while both identity and utility concerns are important drivers of euro support both before and during the crisis, there is a significant shift in the balance of these determinants. This amounts to an increase in the differences in the predicted probabilities of euro support produced by perceived membership benefits and institutional benefits of about 4 -7 percentage points over the course of the crisis, and a similar decrease in the differences in euro support due to identity. This is quite a substantial shift over a relatively short time period of four to eight years, and it is comparable in magnitude to those reported in existing studies on changing determinants of public support for European integration (see Hakhverdian et al. [2013]). Therefore, this analysis has demonstrated that while a simple glance at aggregate-level support for the euro inside the euro area reveals little change over the past eight years, a closer look at individual-level data demonstrates a significant shift in the factors shaping support. Inside the euro area the balance between rational cost-benefit considerations and identity factors has changed: economic benefits and institutional capacity now matter more, whereas identity matters less. In contrast, the determinants of support for the euro have remained stable outside the euro area.

CONCLUSION
Public opinion and voters play only a marginal role in classic theories of European integration. According to the dominant theories, the process of integration has been considered the remit of political and economic élites and of little interest to ordinary citizens. However, the euro area crisis has presented the EU with its most acute challenge to date, as it has pitted creditor states against debtor states, brought the negative consequences of monetary integration into focus and threatened the very survival of the new currency. As a result, Europe has become a more salient and divisive issue than ever before and European leaders can no longer rely on a 'permissive consensus'. In this contribution, we have addressed the important question of how the crisis has affected the nature of attitudes towards the euro.
Our findings show that despite the severity of the crisis, support for the euro inside the euro area has remained high, whereas it has declined significantly outside the euro area. Moreover, citizens have become more aware of these costs due the heightened salience of the economic consequences of the integration process during the crisis. We argue that utility calculations have become more important to euro insiders during the crisis, at the expense of identity heuristics that have come to play a less important role in shaping attitudes. Indeed we find strong support for these propositions in our analysis of cross-national individual-level survey data between 2005 and 2013, which demonstrates that the balance between rational cost -benefit considerations and identity factors has changed inside the euro area, whereas we find virtually no change outside the euro area. Hence, the analysis reveals a growing divide between insiders and outsiders: not only have citizens inside the euro area remained more supportive of the euro than those outside, but they also appear more readily persuaded by utilitarian considerations and arguments than citizens in other EU countries.
These findings contribute to the wider literature on the dynamics of support for European integration. Recent work on European integration, notably the theory of postfunctionalism (Hooghe and Marks 2009), has highlighted that national identity should become increasingly important for attitudes on European integration as the EU becomes politicized in domestic arenas. This is based on the observation that political entrepreneurs, often located on the fringes of the political spectrum, use identity frames to mobilize the integration issue in public discourses. This study suggests that the shifting determinants of EU support may depend on the type of politicization that takes place. Our findings show that as the issue of monetary integration became more salient and citizens more aware of it, identity concerns became less important to citizens, and utilitarian concerns about the EU's institutional effectiveness and benefits of integration became more important. This is not necessarily irreconcilable with the postfunctionalist propositions. Hooghe and Mark's expectation is that identity-based public opinion will be particularly pronounced if the integration issue that is contested has 'opaque economic implications ' and 'transparent communal implications' (2009: 13). However, in contrast to previous periods of high politicization of the EU issue (notably referendums on treaty changes), the euro crisis has emphasized the economic and redistributive implications of integration (Hobolt and Tilley 2014;Kriesi and Grande 2014). Hence, this suggests that when the debate on European integration is focused on economic integration and its consequences, this may instead encourage citizens to think of integration more in terms of economic self-interest and less in terms of their national identity. It is also noteworthy that the increased salience of European integration has been driven not primarily by the issue entrepreneurship of political élites and challenger parties, but rather by an exogenous 'shock' to the EU's economic system in ways that have not only shaped the nature of the public debate, but also had tangible effects on the personal circumstances and future prospects of millions of citizens in the EU. Interestingly, our findings show that the shift in the dynamics of euro support took place before the worst effects of the crisis were felt by ordinary Europeans, and hence can be interpreted as a reaction to the heightened salience of integration and the fear of future repercussions, rather than actual changes in economic circumstances.
The increasing importance of the 'utility logic' over the 'identity logic' may also explain why support for the euro has remained high and stable inside the euro area, despite the severity of the economic crisis. Our results suggest that Europeans generally considered the EU to be more effective than any other institution, including their own national governments, in taking action against the crisis, and inside the euro area the perceived effectiveness of EU institutions played an increasingly important role in keeping support for the euro high. It is worth noting, however, that the shift away from identity-based opinions may be particularly pronounced because we are looking specifically at attitudes towards monetary integration, which arguably has clearer economic implications than other forms of integration, and it is not certain that these findings can be generalized to all types of integration support.
This study nonetheless has several implications for our understanding of European integration. First, it suggests that we should not necessarily expect greater politicization of the integration issue to equal more identity-based attitudes towards integration. Indeed, greater contestation that highlights the distributional consequences of the integration process may instead lead to opinions that are increasingly based on a cost-benefit analysis of the European project. A second and related contribution is that it challenges the prediction that national identities will continue to pose a significant constraint on the integration process in the long term. While our results show that identities matter, they also indicate that major shocks to the system can render them less important, certainly when it comes to attitudes towards economic integration. So far, the integration issue has been mobilized primarily by domestic political entrepreneurs emphasizing identity-based Eurosceptic concerns. Yet, our findings suggest that the public may also be receptive to the mobilization of the integration issue on the basis of utilitybased arguments. Hence, while we are unlikely to return to the days where the European political élites could safely ignore public opinion, this study has illustrated that greater public contestation does not necessarily equal a public veto. Instead, the crisis demonstrates that public opinion on integration might be more dynamic and responsive to the changing nature of the integration process than stylized theories predict. This should be reflected in future studies as we continue to advance our models of public opinion formation on European integration by taking seriously the impact of the political and economic context.