Juncker's Curse? Identity, Interest, and Public Support for the Integration of Core State Powers

In this study we analysed the patterns and covariates of public support for the European integration of core state powers based on an original new survey. We found considerable variation across integration instruments, member states and policy issues. Horizontal transfers are supported more than vertical capacity building; member states from the EU's South&#8208;East are more&#160;supportive than states from the North&#8208;West; and support increases from debt relief to unemployment assistance, sharing the burdens of refugees, and military defence to disaster aid. Identity is a strong and fairly consistent predictor for individual variations in support. The association with respondents&#8217; interest is less consistent, but can be quite strong with respect to specific policy issues such as debt and unemployment. Overall, support for the integration of core state powers is higher and more variable than expected. This suggests there is considerable room for political agency rather than a general constraining dissensus.


I. A Post-functional Predicament?
The recent crises of the EU have fuelled a functional demand for the integration of core state powers (CSPs) (Genschel and Jachtenfuchs, 2018). More and stricter rules of national conduct no longer seem enough to ensure the viability of the EU. What appears to be required is instead the sharing among member states of the key resources of sovereign government (money, coercion, and administration): • The Euro crisis highlighted the need for a common fiscal backstop (such as, common debt, common taxes, and a common budget) to quell self-fulfilling liquidity crises in the Eurozone. • In the wake of the Euro crisis, (youth) unemployment in some member states rose to levels not seen since the 1930s, triggering calls for a common EU unemployment insurance. • The refugee crisis led to proposals for a joint refugee relocation scheme, the transformation of Frontex into an independent European border police and the creation of a common administration of asylum applicants and policies. †Earlier versions of this article were presented in Berlin (October 2018), Amsterdam (December 2018), Brussels (February • Increased Russion military assertiveness, the prospects of a US withdrawal from NATO and the erosion of the post-Second World War peace order have triggered calls for an integrated European army to increase military security and leverage European power.
Despite the pronounced functional demand, political supply has remained meagre at best. Even during the worst time of the Euro crisis, fiscal capacity building through the European Stability Mechanism or the Outright Monetary Transactions programme of the European Central Bank was highly contested, and invariably late. During the refugee crises a joint refugee relocation scheme was agreed but it was not consistently implemented. Plans for strengthening Frontex were mutilated by sovereignty concerns. Emmanuel Macron's call for a European army or Olaf Scholz' proposal for a common European unemployment scheme retain a dreamy pie in the sky quality.
Why is the integration of CSPs so difficult, even in the face of imminent collective disaster? Perhaps the former President of the European Commission, Jean-Claude Juncker, has the answer: 'We all know what to do but we just don't know how to get re-elected once we have done it' (quoted in Buti et al., 2009, p. 65). In principle, the governments of member states can agree on effective European problem-solving, including the integration of CSPs; in practice, they are hindered by their voters. In theoretical terms, neofunctionalism and liberal intergovernmentalism theories hold that political elites and economic interest groups support integration if it promises collective gains in situations of asymmetric interdependence (Moravcsik, 2018;Sandholtz and Stone Sweet, 2012). Post-functional theorists warn, however, that '(m)ass politics trumps interest group politics when both come into play' (Hooghe and Marks, 2009, p. 18). As political entrepreneurs, especially challenger parties from the radical populist right (RPR), mobilize the parochial identities and interests of voters', incumbent parties increasingly 'worry about the electoral consequences of their European policies' (Hooghe and Marks, 2009, p. 21). Their willingness to compromise on EU issues decreases as those become politicized (Hutter and Grande, 2014;Hutter et al., 2016;de Wilde et al., 2016). This increases the risk that integration will fail even when it is efficient and useful.
Does Juncker's curse explain the undersupply of CSP integration? To find out, we need to know what voters think. This information is hard to get from existing surveys such as the Eurobarometer or the European election study because they do not contain specific questions on attitudes to the integration of CPSs. In this article we present new survey evidence that helps fill this gap. It is based on a 2018 YouGov poll that asks specifically whether respondents support or oppose ceding national fiscal, coercive, or administrative resources to other member states or to EU institutions. This survey thus offers a better measure of our dependent variable (voter attitudes to the integration of CSPs) than alternative sources discussed below.
In the next section we develop our analytical framework for presenting and interrogating the data. We define CSPs and discuss interest-based and identity-based reasons why voters may oppose or support their integration. In section III we present our survey and empirical approach. In sections IV and V we present the evidence. We found that public support for the integration of CSPs was positive overall, though the level varies markedly across countries, individuals, issues and instruments. We also found that identity is a stronger predictor than interest for individual variations, but not for variations at the country level. Finally, we found that the share of 'don't knows' is often but not consistently high, suggesting considerable uncertainty and ambiguity about the desirability of the integration of CSPs among voters. In section VI we conclude that there may be more room for political agency than Juncker's curse suggests.

II. Why Should Voters Care?
CSPs are the action resources derived from the state's twin monopoly of legitimate coercion and taxation Jachtenfuchs, 2016, p. 43, 2018, p. 181). They include its coercive capacity (the military, police forces, and border patrols), its fiscal capacity (money, taxes, and debt), and the administrative capacity (people and organization) to implement and enforce public laws and policies. CSPs constitute the state: states without sufficient coercive, fiscal, and administrative resources are failed states at best (Besley and Persson, 2013); institutions possessing such resources are states de facto, if not always de jure (for example, Taiwan).
After the failure of the European defence community treaty in 1954 the EU tried to steer clear of CSPs, focusing on market integration instead. Yet, the completion of the single market, the introduction of the euro and the abolition of border controls as well as southern and eastern enlargement have pushed the EU into CSPs, nevertheless. They have created new interdependencies that cannot be managed by market regulation alone. The issue since then has been no longer only the will of member-state governments to cooperate and avoid negative externalities (which could be addressed by stricter rules and better enforcement) but also their ability to do so (for this distinction, see Chayes and Chayes, 1993). Eurozone countries going into deficit to bail out banks or to deal with massive unemployment, member states confronted with many refugees, or states feeling threatened by Russia do not need more rules but more resourcesmoney, force, or bureaucrats. The cohesion and viability of the EU then depends on the sharing of financial, coercive, or administrative resources (Ferrera and Burelli, 2019;Genschel and Jachtenfuchs, 2018;Schelkle, 2017). This can happen in two ways: • Through horizontal transfers, by which one state sends money, credit guarantees, police agents, administrators, or troops to another state that lacks them (obviously, the same can happen between groups of states). No EU resources are involved, but EU institutions regulate, coordinate, or facilitate the transfers. The first Greek rescue package is an example. • Through vertical capacity building, which creates genuinely European CSPs independent from member states (a European monetary fund, a European border police, or a European army), thus relieving member states unable to mobilize sufficient powers on their own of the burden of doing so. An example of this is the outright monetary transactions programme of the European Central Bank.
Why should mass publics oppose (or support) horizontal transfers or vertical capacity building? Intuitively, there are two main reasons: fear of material loss and fear of ideational loss. As the large literature on public opinion towards the EU suggests (de Vries, 2018;Hobolt and de Vries, 2016;Jupille and Leblang, 2007) people may resist integration because they believe that it implies more material costs than benefits to their own state (and hence presumably also to them individually). Alternatively, they may resist integration because they perceive it as a threat to their national identity and sense of belonging.
The integration of CSPs is thus likely to have high salience both for people's distributive interests and their sense of identity (Genschel and Jachtenfuchs, 2018;Hobolt and de Vries, 2016;Jupille and Leblang, 2007;de Vries, 2018;Winzen and Schimmelfennig, 2016). Take interests first. As fiscal, coercive and administrative resources are very costly to mobilize and maintain, integrating them immediately raises the issue who pays and who benefits, who gains and who loses. Given the size and heterogeneity of EU member states it is unlikely that national net balances will always be completely even. Structural asymmetries may turn the horizontal or vertical sharing of CSPs into a permanent redistribution mechanism from fiscally sound to fiscally weak member states, from competitive economies to less competitive ones, or from states that are protected by their geography from refugee inflows to those exposed to inflows and so on. Also, there is a moral hazard involved in sharing or pooling CSPs: some member states may be tempted to free ride on resources provided by others. At the same time, however, the integration of CSPs may also be a source of material gain. People may believe, for instance, that their home state will mostly be at the receiving end of horizontal transfers or that vertical capacity building will unleash economies of scale with benefits for all member states. These arguments lead to the following expectation: Interest conjecture: Individuals will oppose the integration of CSPs if, and to the extent that, they expect a net material loss for their own state; they will support integration if, and to the extent that, they expect a net material gain.
Turning to identity next, CSPs are intimately linked to people's sense of collective identity, that is, to 'that part of "me" that belongs to a larger "we"' (Risse, 2010, p. 22). CSPs constitute the state that defines the political community; they provide that community with the essential means of self-government; they standardize experiences across the community by imposing the same money, the same taxes, and the same rules and regulations on all citizens by the same administration and enforcement regime. All this links the integration of CSPs tightly to fears (or hopes) of the demise of the nation state and the emergence of a European super-state. Individuals with a strong and exclusive sense of national identity will perceive this as a threat that they oppose.
Individuals with more inclusive identities will be more sanguine about integration. In some cases, such as post-Franco Spain, people may even perceive European integration to be a boon to national identity because it confirms the normalcy and modernity of their own nation (Diez Medrano, 2003). Also, despite the prevalence of national identity, many people feel European; some of them even exclusively so (Diez Medrano, 2003;Fligstein, 2008;Risse, 2010). To the extent that they do, they may support the integration of CSPs as a way to realize that identity. A strong collective identity facilitates solidarity and mitigates fears of material loss. Purportedly, a widespread sense of German identity allowed the German federal government to transfer roughly six per cent of GDP annually to East Germany without major protest in West Germany during the 1990s (Scharpf, 1999, p. 9, note 3). These arguments lead to the following expectation: Identity conjecture: Individuals will oppose the integration of CSPs if, and to the extent that, they hold exclusively national identities. They will support integration if, and to the extent that, their sense of national identity is complemented or even dominated by European identification.
Interest and identity are analytically distinct determinants of individual attitudes. This makes them useful guides for a systematic comparison of attitudes to the integration of CSPs. Empirically, however, they often interact in important ways: 'who we are influences what we want' and vice versa (Abdelal et al., 2006, p. 698; see also Cram, 2012, p. 75;Kohli, 2000, p. 118;Kuhn and Nicoli, 2020). People with an exclusively national identity may perceive the material costs and benefits of integration very differently from those with inclusive European identities; cost-benefit perceptions, in turn, may shape whether people hold national or European identities. For instance, Hobolt and Tilley (2014) find that citizen's attribution of responsibility in the EU is influenced by groupserving biases. Moreover, there is evidence that, in Britain, committed Leavers have a very different view of reality from that of committed Remainers (Curtice, 2017, p. 31;Hobolt, 2016Hobolt, , p. 1270. This needs to be taken into account when we assess the relative power of the interest and the identity conjecture below.

III. Data and Empirical Approach
Comparative data on attitudes to the integration of CSPs is scarce. In the wake of the financial crisis, various studies have analysed attitudes to financial assistance. Some of these used very specific measures, including support for the European Stability Mechanism in Germany (Bechtel et al., 2014) and votes in bailout referenda in Greece and Iceland (Walter et al., 2018;Curtis et al., 2014). Others relied on general indicators of financial solidarity (Stoeckel and Kuhn, 2018;Verhaegen, 2018) or support for joint economic governance (Kuhn and Stoeckel, 2014). None of these studies included a comparison across issues (such as attitudes to the integration of CSPs for purposes other than debt relief and financial assistance) or instruments (such as horizontal transfers between states versus vertical capacity building at the EU level). Existing surveys (for example, Eurobarometer or the European election study) simply lack the relevant questions (see also Kleider and Stoeckel, 2019, p. 11).
A new online survey by YouGov helps to fill this gap. The survey was fielded in April 2018. It covers 11 EU member states from Western Europe (Germany, France, and the UK), the Nordic region (Denmark, Finland, and Sweden), Southern Europe (Greece, Italy, and Spain) and Eastern Europe (Lithuania and Poland). It is broadly representative of the voting age population on basic socio-demographic variables (age, education, and gender), and the total sample size was 11,284 respondents. Table 1 summarizes the relevant questions that we used in this article (for summary statistics, see Supporting Information Table S1).
The first four questions in Table 1 serve as measures of our dependent variable: attitudes to the integration of CSPs. Importantly, the questions cover support for or opposition to different instruments of CSP integration (horizontal transfers and vertical capacity building), not just integration in general. They also examine attitudes on different issues that might be addressed through these instruments (horizontal transfers for debt relief, unemployment insurance, receiving refugees, military assistance, or natural disaster relief and vertical capacity building to create a joint EU army or increase the EU budget).
The last three questions in Table 1 are proxies for our independent variables, identity and interest. The question on the net financial position is a rough measure of respondents' expectations of the net gains or losses associated with integration for their member state. In line with the interest conjecture, we expect respondents who perceive their own Thinking about the money that is spent by the EU and the money that is spent by member states, which of the list below best reflects your view?
Support, more money should be raised and spent by the EU, and less by the member states; oppose, more money should be raised and spent by the member states, and less by the EU (oppose); neither, the current balance is about right (neither); don't know member state as a net contributor to be less supportive of the integration of CSPs than those who perceive their home state as a net recipient. The variable for radical populist right (PRP) voters measures national identity at the individual level (see Table S2 for information about our coding of RPR parties). In line with the identity conjecture, we expect RPR voters to be less supportive of the integration of CSPs because they have a stronger and more exclusive sense of national identity. To be sure, the RPR vote is not an ideal measure of identity because voters have various reasons beyond identity to vote for RPR parties (Rooduijn, 2018). Yet there is widespread agreement that the mobilization of nationalist feelings is key to the electoral success of RPR parties (Hooghe and Marks, 2009;Meijers, 2017;de Vries and Edwards, 2009). At the country level, we rely on the Moreno question to assess the exclusiveness or inclusiveness of respondents' identity. In contrast to the other questions, the Moreno question is not included in the YouGov survey. We drew it from the Standard Eurobarometer 89 (Spring 2018). Our empirical analysis proceeds in two steps. In section IV we map the dependent variable. We compare support and opposition for the integration of CSPs across issues (debt relief, unemployment insurance, refugee assistance, natural disaster, or military attack), instruments (horizontal transfers and vertical capacity building), and countries. In section V we explore how well interest and the identity conjectures account for the observed variance in support. We investigate whether country differences in support are associated with country differences in our measures of material interest and collective identity. Then we explore individual differences across issues and instruments: do voters with an exclusive national identity differ in their attitudes to the integration of CSPs from other voters? How does the perceived net position of the own member state vis-à-vis the rest of the EU affect individual attitudes?
IV. Patterns of Support: Variation across Issues, Instruments, and Countries

Variance across Issues and Instruments
Two survey items measure support for horizontal transfers of CSPs: would respondents support or oppose financial transfers to member states suffering from unsustainable debt, high unemployment, large inflows of refugees, or natural disaster? And would they support military assistance to member states under foreign attack? Two other items gauge support for vertical capacity building: are respondents in favour of an integrated European army? Do they think the EU should raise and spend more money and the member states less? Figure 1 reports the main results (see Genschel and Hemerijck, 2018).
Consider horizontal transfers first. The level of support varies strongly by issue. While nearly 80 per cent of respondents, on average, support assistance to member states in case of natural disaster, fewer than 40 per cent are in favour of helping over-indebted member states. Support is generally higher than opposition (if only by a slight margin for debt relief). It is inversely related to both opposition and uncertainty ('don't know'): high support goes together with low opposition and with low uncertainty about the desirability of horizontal transfers. While support for disaster relief seems clear-cut, support for debt relief is apparently more ambiguous.
In terms of vertical capacity building support is generally lower than for horizontal transfers. While a majority of respondents supports the creation of an integrated European army, only a minority favours fiscal capacity building. The share of respondents wanting to expand EU fiscal capacity (support) is smaller than the share of people wanting either to shrink it (oppose) or to keep it at current levels (neither). 1 Uncertainty ('don't know') about the desirability of fiscal capacity is much higher than uncertainty about the desirability of EU military capacity.
Overall, public support for the integration of CSPs is higher than one would perhaps have expected after the acrimonious conflicts over burden-sharing during the eurozone and refugee crises. Yet, the level of support varies strongly by instrument and issue area. Ironically, support is highest for issues that either have low problem salience (natural disasters) or are beyond the EU's current remit (military security is still largely in the hands of NATO). Support tends to be low, by contrast, for issues of immediate and pressing importance to the EU: debt and unemployment. Only in the case of refugee assistance did we find fairly robust support for an issue of high policy relevance. However, the European averages reported in Figure 1 may mask fundamental cross-national differences in attitudes.

Country Variance
Figure 2 maps respondents' average net support (that is, support minus opposition) for horizontal transfers and vertical capacity building by country. It shows low cross-country variation in issue rankings. Respondents across all member states have fairly similar intuitions on the relative desirability of horizontal transfers. Disaster relief ranks highest everywhere; debt relief ranks lowest almost everywhere; and the ranking of the other issues follows roughly the order suggested by Figure 1. Cross-national variation is even lower for vertical capacity building. Average net support for an integrated European army ranks higher in all member states than net support for an expansion of EU fiscal capacity.
1 As Table 1 shows, the YouGov survey presents financial capacity building as a trade-off between EU capacity and member state capacity. By contrast, the question on military capacity building implies no such trade-off. Respondents are simply asked for their opposition or support for a European army without any national level quid pro quo involved. This difference may reduce the comparability of results and bias the findings.

Figure 1: Public attitudes towards the integration of core state powers by instrument and issue.
Juncker's Curse? Identity, Interest, and Public Support for the Integration of Core State Powers † Despite the similarities in ranking, there are considerable differences in levels of support. Starting with horizontal transfers, we found little variation in support for disaster relief. Net average support is high and positive in all countries. Cross-national variation is more pronounced on refugee inflows and military attack, although net support for transfers on these issues is positive in all countries. Finally, cross-national variation is very high for unemployment and debt relief. While there is positive net support for these transfers in the southern and eastern country group (Greece, Spain, Italy, Lithuania, and Poland), net support is negative in a northern and western group (Germany, Denmark, Great Britain, Finland, France, and Sweden). The differences are stark: while support for debt relief is almost at the level of disaster relief in Greece, there is a huge gap between both issues in Sweden.
Compared with horizontal transfers on debt and unemployment, the cross-national variation in support for European fiscal and military capacity building is surprisingly muted. To be sure, there is net support for both types of capacity building in Greece and there is net opposition to both in Denmark. Yet, the gap between these two extreme countries is lower, and the level of cross-country agreement is higher than in the transfer case. Net support for the European army is generally positive (except in Britain and Denmark). Net support for expanding EU fiscal capacity is generally negative (except in Greece, Germany, and Spain). The north-west, south-east divide is less clear-cut because support in Germany and France is closer to southern and eastern than to northern and western levels.
In conclusion, we found little cross-country variance in the rank order of support for different types of horizontal transfers or vertical capacity building. Yet, we found considerable cross-national variation in the level of support, especially for horizontal transfers for debt relief and unemployment and, to a lesser extent, for vertical capacity building for military and financial purposes. Support tends to be higher in southern and eastern than in northern and western EU member states.

V. Correlates of Support: Interest and Identity
How well do our two conjectures on interest and identity account for variance in attitudes to the integration of CSPs? We investigate this question first at the country and then at the individual level.

Interest and Identity at the Country Level
Are country differences in net support associated with country differences in interest or identity? As a first step, we plotted support for horizontal transfers (Figure 3) and vertical capacity building (Figure 4) against our measures of interest and identity. Our interest indicator is the national average of respondents' replies to the net contributor question in the YouGov Survey (Table 1); our identity indicator is the national average on the Moreno question (see also  Figure 3 plots support for horizontal transfers for two issues: debt relief and refugee assistance (see Figure S1 for similar plots for unemployment relief, military assistance, and disaster relief). It shows a strong association between distributive interest and support for transfers in the case of debt relief: people favour horizontal transfers when they think other member states will pay for them (that is, when their own country is perceived as a net recipient); they favour transfers much less when they think their own member state will have to pay (that is, when their own country is perceived as a net contributor). In short, attitudes are strongly related to distributive interests as the story of the creditordebtor cleavage during the eurozone crisis would suggest. The same applies to unemployment relief (see Figure S1).
Yet, for refugee assistance we found essentially no relationship. The same applies to horizontal military assistance and disaster relief (see Figure S1). Distributive interests apparently do not matter in these issues. Alternative measures of interest (such as the number of asylum applications per capita in the case of refugee assistance or the distance from Moscow in the case of military assistance; see Figure S2) do not change the picture: the association between interests and support for horizontal transfers remains weak and insignificant. What about identity?  Figure 3 shows essentially no association between support for horizontal transfers and European identity. Support is high in countries where European identity is generally weak (for example, Greece) but also in countries where European identity is strong (for example, Spain). The same pattern holds in unemployment relief, military assistance and disaster relief (see Figure S1). In the Supporting Information we also use an alternative measure stressing the emotional (rather than cognitive) dimension of collective identity (Cram, 2012; see also the introduction to this Special Issue). This yields an even more surprising pattern: countries in which people feel happier about living in the EU tend to be less supportive of horizontal transfers (see Figure S3). Figure 4 plots support for vertical capacity building against our indicators of interest and identity. The pattern is broadly similar to Figure 3. Support for military capacity is associated with distributive interest (see Figure S2 for an alternative measure of interest). It decreases as the proportion of people viewing their own country as a net contributor increases. Yet, surprisingly, support for financial capacity is essentially unrelated to interest. Support also does not vary with identity. A higher share of people with a European identification does not significantly increase support for a EU army or EU fiscal capacity, whether or not we use the Moreno question (Figure 4) or the happiness measure of identity (see Figure S4).
In conclusion, we found mixed support for the interest conjecture. In some instances (that is, horizontal debt relief as well as horizontal unemployment assistance and vertical military capacity building), attitudes to European CSPs are indeed correlated with distributive interests, as this conjecture suggests. Yet, we found no such correlation for horizontal refugee assistance, military assistance, disaster relief, and vertical fiscal capacity building. We found no support for the identity conjecture. Country-level differences in European identification are essentially unrelated to country-level differences in attitudes to the integration of CSPs. To be sure, this non-finding could simply reflect our small sample size of only 11 states. Hence, we reproduced our analysis with 28 member states for the one Eurobarometer item that explicitly considers the integration of CSPs: support for a European army. The results are shown in Figure S7 and they look similar to that in Figure 4: European identification and support for a European army are essentially unrelated.

Interest and Identity at the Individual Level
Obviously, a non-finding at the country level does not imply that identity does not matter at the individual level. We performed a simple logit regression to avoid the ecological fallacy. The dependent variable was individual support for horizontal transfers or vertical capacity building (questions 1-4 in Table 1). The key independent variables were constructed from the net contributor and the RPR vote items, respectively (Table 1). The former provide our indicator of interest. It is a categorical variable with three values: net contributor, net recipient, or balance depending on how respondents perceives their home country's net balance with the rest of the EU. The latter is our identity measure. It is a dummy variable that takes the value of 1 if the respondent has voted for an RPR party (indicating an exclusive national identity) and 0 otherwise. Finally, we entered three common socio-demographic control variables from the YouGov survey: age, educational attainment, and gender.
As Table 2 shows, RPR voters are significantly less likely than other voters to support the integration of CSPs. The regression coefficient for RPR voters is large and statistically significant for all issues, and instruments, and it is particularly large for refugee assistance. Obviously, identity matters at the individual level. Distributive interests also matter, but less consistently: the perceived net position of one's own state is a much weaker predictor of individual variations. Respondents who think of their country as either a net recipient of EU funds or as having a balanced net position vis-à-vis the EU tend to offer more support for horizontal transfers than voters who think of their country as a net contributor (the reference group). Notable exceptions are natural disasters and, to a lesser extent, military attack and financial capacity building. Interestingly, both identity and interest are less strongly and consistently associated with attitudes to vertical capacity building than attitudes to horizontal transfers. Table 2 looks only at main effects. Yet identity and interest may interact in important ways (see section II). First, exclusive nationalists may have a higher propensity to view their own country as a victim of EU exploitation. Indeed, our data suggest that RPR voters are somewhat more likely to perceive their country as a net contributor to the EU than other voters (tables S11 and S12, Figure S9) Secondly, exclusive nationalists may feel more strongly about the political implications of their country's net position. Conceivably, RPR voters oppose integration particularly strongly if they fear that their own country will have to pay for it and support integration strongly if they believe it will benefit their own nation. Finally, people who conceive of their home state as a net contributor may more easily adopt an exclusive national identity to justify their interest-based opposition. To test these intuitions, we added interaction effects to the regression models (Table S3). As interaction effects are difficult to interpret by coefficients alone, we plotted them for selected scenarios ( Figure 5; see Figure S5 for plots of the other scenarios).
Three observations stand out. Firstly, RPR voters tend to support horizontal transfers and vertical capacity building less than other voters. Secondly, the size of the support gap varies by issue and instrument: it is much larger for the refugee issue than for the debt issue, and it is larger for horizontal transfers than for vertical capacity building. Thirdly, there is a weak interaction effect. While the gap in support between RPR and other voters is often significant if their own country is perceived as net contributor or as having a balanced position, it is insignificant if their own country is perceived as a net recipient (except for the refugee issue). This suggests that RPR voters hold slightly more opportunistic attitudes to the integration of CSPs than other voters.
On the individual level, we thus found clear support for the identity conjecture. The effect of an exclusive national identity (proxied through the RPR vote) on support for the integration of CSPs is consistently negative, but the size of the effect varies across issues and instruments. It also varies in terms of the perceived net position of the own country: if the home country is perceived as a net recipient of EU funds (that is, if integration brings material benefit), the attitudes of RPR voters converge towards those of other voters (except for the refugee issue). In general, however, support for the interest conjecture is mixed. Distributive interests are a less powerful predictor of individual attitudes than identity. Only support for horizontal debt relief and unemployment assistance is strongly related to the perceived distributive position of the home country. Finally, our   indicators of identity and interest are less consistently associated with support for vertical capacity building than to support for horizontal transfers. 2

VI. Juncker Relax!
Juncker's curse has become a folk theorem of practitioners and scholars alike. In principle, it claims, governments are willing and able to strike efficient and effective bargains on European integration, including the integration of CSPs. In practice, they are hindered by voters. Liesbet Hooghe and Gary Marks have turned Juncker's curse into a post-functionalist theory of EU integration (Hooghe and Marks, 2009). It starts from a critique of the functional optimism of neo-functionalist and liberal inter-governmentalist scholars who believe in the ability of EU elites to strike efficient bargains under conditions of asymmetric interdependence and largely unencumbered by mass politics. Its main contribution is to theorize the potentially inhibiting role of mass publics, party politics and collective identity on EU policies and institutions. It insists that voters are not just a nuisance factor for EU elites but may shape integration outcomes in major, if not necessarily salutary, ways (see also Hooghe and Marks, 2018).
In this article we assessed the extent of a post-functionalist constraining dissensus on the integration of CSPs. Although the EU's recent string of major crises was closely related to CSPs, there is scarce information on what voters think about their integration. We present here new survey evidence to help close this gap. Our analysis reveals a complex picture of public attitudes: • Firstly, there is no general constraining dissensus with respect to CSPs. Overall, support for them outweighs opposition to them. Public opposition is limited to some issues, some countries and some voters. It is driven by both interest and identity. An exclusive national identity is generally associated with lower support for the integration of CSPs. Interest kicks in selectively when distributive implications are stark and obvious (debt and unemployment relief). The strongest resistance comes from those who fear having to pay the bill (respondents in northern and western countries over financial transfers for debt and unemployment relief) and those who dislike refugees (RPR voters). • Secondly, there is explicit support for CSP integration on some issues and instruments: even RPR voters are more likely to support than oppose horizontal military solidarity and the creation of a European army, and even in Germany there is net support for more fiscal capacity. • Finally, opposition to the integration of CSPs is tempered by high cognitive uncertainty: in our data low support is generally associated with high uncertainty, that is, high shares of respondents who do not know whether they support or oppose integration. This is important because where voters are uncertain political leaders can cue themnot only those from the RPR but also from the rest of the political spectrum.
Evidently, there are limits to our analysis. Some are due to data restrictions. Perhaps most importantly, our measures of identity and interest are very basic and could be improved. Also, the limited number of countries in the YouGov survey prevented us from analysing the interaction between individual indicators of subjective interest and identity and country-level indicators of objective structure and opportunity systematically in a multilevel framework (see Kleider and Stoeckel, 2019, for such an analysis). Other limitations are due to findings for which we have no good explanation. Perhaps most surprising to us was how little work the distinction between horizontal transfers and vertical capacity building seems to do for the empirical analysis. To be sure, the general level of support for the former is higher than for the latter. Yet the variance in support is also higher for horizontal transfers. This is puzzling, given that the creation of vertical capacities is likely to be more durable and intrusive and more closely associated with European state-making than horizontal transfers.
In conclusion, governments should listen carefully to what voters want. But they should also be careful not to misread public opinion (Bremer, 2019). There is no reason to do something that voters did not ask for, or not to do something that voters would not object to. If voters are uncertain what to think, there is nothing wrong with telling them. Hence, our analysis provides some guarded hope for Jean-Claude Juncker and the new President of the European Commission Ursula von der Leyen. If EU elites do indeed know what to do, they should go ahead and do it. The electoral risks from intense nationalist minorities are obvious. Yet stalling is also risky. Large majorities want to keep the EU. If this requires more integration of CSPs they may actually support the political elites who give them the choice to opt for it (Ferrera and Burelli, 2019, p. 106).

Supporting Information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Table S1. Summary statistics of key independent and dependent variables from the YouGov survey Table S2. Parties coded as radical populist right Table S3. Support for horizontal transfers and vertical capacity building in all countries with interaction effects (logit regression) Table S4. Support for horizontal transfers and vertical capacity building in all countries (logit regression, without survey weights) Table S5. Support for horizontal transfers and vertical capacity building in all countries with interaction effects (logit regression, without survey weights) Table S6. Support for horizontal transfers and vertical capacity building in western Europe (logit regression) Table S7. Support for horizontal transfers and vertical capacity building in western Europe with interaction effects (logit regression) Table S8. Opposition to horizontal transfers and vertical capacity building in all countries (logit regression) Table S9. Opposition to horizontal transfers and vertical capacity building in all cuntries with interaction effects (logit regression) Table S10. Support for a European army (logit regression) Table S11. Perceived interest (net financial position) by identity (vote for the radical populist right) Table S12. Perceived interest and identity (logit regression) Figure S1. Average net support for horizontal transfers by interest and identity (other scenarios) Figure S2. Average net support for horizontal transfers and vertical capacity building with alternative variables for interest. Figure S3. Average net support for horizontal transfers by happiness about living in the EU Figure S4. Average net support for vertical capacity building by happiness about living in the EU Figure S5. Predicted probability for supporting the integration of core state powers by identity and interest (other scenarios) Figure S6. Predicted probability for opposing the integration of core state powers by identity and interest Figure S7. Support for a European army by identity Figure S8. Predicted probability for supporting a European army by identity (Moreno question) Figure S9. Predicted probability for perceiving one's country as net creditor by identity