The politicisation of European integration in domestic election campaigns

Abstract How strongly is European integration being politicised in election campaigns, and what explains why a party chooses to emphasise Europe or, by contrast, remains silent about it? This article provides a systematic assessment of the salience of European integration in domestic election campaigns, tracking its development from the 1990s to the 2000s across six Western European countries based on media content analysis data. The findings show that the salience of Europe in election campaigns is actually rather limited when put into perspective by benchmarking it against other political issues. Moreover, ideological determinants are crucial in explaining European integration issue-emphasis. In particular, the more culturally conservative a party, the stronger its emphasis on Europe; the impact of the economic left–right divide, by contrast, is weaker and more ambiguous. However, Europe remains in the shadow of its twin issue, immigration, which shares a similar issue-emphasis pattern yet is more attractive to these culturally conservative parties.

the radical left have 'smelled blood' and have not passed up on this opportunity (Hooghe and Marks 2009: 21). Hence, the view of European integration as being a successfully and strongly politicised issue has come to dominate the scholarly debate. Only a few dissenting voices critically object that the extent of Europe's politicisation is vastly overestimated. For them, European integration is, and will remain, a marginal issue in domestic politics -a 'giant fast asleep' and unlikely to fundamentally reshape traditional patterns of political conflict and competition (Green-Pedersen 2010; also see Mair 2001). However, the truth is that we know surprisingly little about the actual extent of the politicisation of Europe. So far, most scholars have simply provided some anecdotal evidence to support their view -e.g. pointing to individual spectacular events such as the referendum defeats of the EU Constitutional Treaty or dramatic episodes of the Eurozone crisis. And even with systematic data at hand, there is still the question of how to properly evaluate the politicisation of Europe -i.e. to decide whether the glass is half full or half empty.
Against this backdrop, the present article sets out to provide a systematic assessment of the long-term politicisation of European integration in domestic election campaigns. Domestic elections constitute the most significant political events in modern representative democracies. And while the politicisation of Europe is not restricted to the partisan arena, the politicisation of any new political issue -if actually relevant -is sooner or later becoming manifest in this setting. Moreover, unlike during European referendums or in front of policy-specific publics, here European integration meets stiff competition from a potentially unlimited number of pressing political issues. This allows for the politicisation of Europe to be put into perspective by benchmarking the salience of European integration against other political issues. In order to do this, the empirical analysis relies on data from a quantitative mass-media content analysis of election campaigns from the 1990s until the 2000s (with the early 1970s serving as an additional reference point) in six Western European countries (Austria, France, Germany, Netherlands, Switzerland, and the UK).
While relevant on its own, the descriptive question of the actual extent of the politicisation of Europe goes deeper as it is related to the question of what actually drives it. In fact, the conflicting scholarly expectations about the extent of the politicisation of Europe are also motivated by differing views about the crucial underlying dynamics. Hence, in a second step, the article tests these competing views about the key factors driving politicisation by looking at the reasons a party chooses to talk frequently about European integration in election campaigns or, by contrast, remains silent about it (European integration issue-emphasis). The above-mentioned sceptical voices claim that the -in their view greatly limited -politicisation of Europe is mainly motivated by strategic considerations of competing parties in opposition or at the fringes of the political spectrum. Prominent scholars of the opposite camp argue, in contrast, that the alleged extensive politicisation of Europe is spurred by this issue's connection to highly salient, more basic ideological lines of conflict (e.g. Hooghe and Marks 2009;Kriesi 2007).
My findings suggest that, in line with the latter camp, the politicisation of Europe is indeed significantly shaped by fundamental political divides. Most strikingly, the closer a party is to the culturally conservative pole, the more likely it is to emphasise European integration. The impact of the economic left-right divide is also observable, though more ambiguous. However, contrary to the claims advocated by many of these same scholars, this ideological issue-emphasis pattern goes along with the severely limited salience of European integration. Certainly, the salience of Europe has increased since the 1970s. Yet, compared to other political issues, the level of salience is still moderate. In particular, Europe remains in the shadow of immigration, an issue that is also part and parcel of an emerging new cleavage between the winners and losers of globalisation (Kriesi et al. 2006(Kriesi et al. , 2008(Kriesi et al. , 2012 and which is also strongly culturally driven. Further, the findings provide no support for the competing hypothesis that fringe and opposition parties make use of European integration for purely strategic reasons.
The remainder of the article is structured as follows: The next section outlines the theoretical framework and, based on the existing literature, derives potential scenarios regarding the extent and the underlying factors that drive the politicisation of Europe. Following a section on methods and data, the empirical part presents the results, first on the overall salience of Europe in election campaigns, then regarding the factors that determine how strongly individual parties emphasise European integration. The article concludes with a discussion of the key findings and their further implications.

The role of European integration in domestic election campaigns
The question of whether the politicisation of European integration is either substantial or marginal calls first and foremost for a proper description of the actual extent of politicisation. However, as mentioned above, conflicting scholarly expectations about the extent of the politicisation are motivated by differing views as to which underlying dynamics actually drive contestation over Europe. Politicisation does not happen automatically or simply by chance; politicians and their mobilisation strategies are crucial to its success or failure (Hooghe and Marks 2009: 18). In the following, I shall discuss in some detail the key factors that are expected to shape how European integration is being politicised by the parties: fundamental ideological divides and strategic considerations related to party competition.
Among those scholars who argue that the politicisation of Europe by the parties primarily follows an ideological pattern, there is some disagreement on how exactly European integration is related to more general lines of political conflict. One can reasonably conceive of the political space in Western Europe as essentially having a two-dimensional structure. Politicians do not compete along a single left-right dimension, but along both an economic and a cultural axis (Hooghe et al. 2004;Inglehart and Flanagan 1987;Kitschelt 1995;Kriesi et al. 2008). The economic left-right axis sees adherents to market intervention and an encompassing welfare state opposed to proponents of market liberalisation and lean government. The cultural axis has changed its meaning over time: with the waning of the religious cleavage, it currently stretches from traditional, authoritarian and nationalist (TAN) values to green, alternative and libertarian (GAL) views. 1 While some authors have claimed that the politicisation of Europe is primarily driven by economic rationales -e.g. adherents of regulated capitalism versus free marketers (Tsebelis and Garrett 2000; also see Leupold 2015) -it has more recently been argued that the issue is strongly attached to the cultural axis, more so than to the economic one (Hooghe and Marks 2009;Hooghe et al. 2004;Kriesi 2007). European integration, in this view, has become a political conflict mainly between GAL supporters and opponents at the TAN pole. The latter include not only fringe parties from the populist radical right, but also mainstream right parties such as the British Conservatives and the German CDU/CSU. Hooghe, Marks and colleagues (2009: 17;2004: 131-2) suggest that this opposition from the TAN pole is motivated by more general concerns about national sovereignty and national identity, which these parties are eager to protect. Kriesi et al. (2012;also see Kriesi 2007) argue similarly, but see opposition to European integration as part of a potent newly emerging globalisation cleavage that pits winners of denationalisation against losers. In their view, this conflict is not exclusively cultural, but also contains an economic dimension related to deregulation and the opening up of previously protected domestic markets. However, the empirical findings by Kriesi et al. suggest that this new cleavage is successfully mobilised by TAN parties only and therefore mainly expressed in cultural terms. At least with respect to European integration, a possible explanation for this might be that opposition on economic grounds can be motivated from the left and from the right as the EU has been active in both market-making and social (re-) regulation. Consequently, one would expect the politicisation of Europe to be primarily culturally driven, although economic motives might play a role as well, albeit a more ambiguous one.
H1: The politicisation of Europe by the parties is shaped by their ideological characteristics, i.e. their location on the economic left-right axis and the cultural TAN-GAL axis of the political space. More specifically, the closer a party is to the culturally conservative TAN pole, the more it is expected to politicise European integration. The impact of the economic left-right axis, by contrast, is expected to be more limited as opposition to Europe here is not clearly concentrated at one pole.
On a related note, one would expect a similar pattern of politicisation for the issue of immigration because it is also supposed to be part of the emerging globalisation cleavage and strongly culturally driven. Hence, it will be instructive to see in the empirical analysis how European integration compares with this twin issue. What distinguishes the two is that European integration is a more complex, multi-faceted issue that is difficult to handle for most politicians as it is a moving target, attitudes to it are often ambivalent, and the framing of what European integration is actually about is highly contested. As a result, the politicisation of Europe, although probably similar in nature, should be less straightforward than the politicisation of immigration.
The alternative view about the nature of the underlying conflict claims that the issue of European integration is orthogonal to existing political divides. In other words, the politicisation of Europe would be largely independent of the traditional programmatic profiles of the parties. At the beginning of the life of an issue, such orthogonality is the norm. Parties on the losing side and new challengers try to come up with fresh, potentially disruptive issues in order to change the existing structure of competition to their advantage (Carmines 1991). It then takes some time for established politicians to incorporate a new issue into their programmatic profiles. In the case of European integration, this might not be feasible as orientations towards such a multi-faceted issue are difficult to fit in. European integration may remain 'orthogonal' -i.e. a single maverick issue defying any stable and coherent ideological linkages (also see Bartolini 2005: 321-2;Steenbergen and Marks 2004). As a result, the issue of Europe may either become marginalised or, by contrast, establish a fundamentally new basic line of conflict (as argued by Hix and Lord 1997: 49-50; also see Benoit and Laver 2006: Chapter 5). In both scenarios, fringe parties, eager to take a Eurosceptic stance simply to signal to voters that they are different from the (pragmatically pro-European) mainstream, and parties in opposition more generally would be expected to mobilise on the issue in order to challenge the governing coalition (Mair 2001;Sitter 2002;Taggart 1998). Based on this reasoning about strategic interaction in party competition, two instrumental hypotheses follow: 2 H2.1: Extreme parties at the fringes of the political spectrum are more likely to politicise Europe than parties which are ideologically moderate.
H2.2: Parties in opposition are more likely to politicise Europe than parties in government.
These expectations regarding the politicisation of Europe will be tested in the empirical section by looking at a crucial element of politicisation, namely the salience of this issue and the parties' efforts at issue-emphasis -i.e. the relative frequency with which European integration is dealt with compared to other issues. The politicisation of an issue is a multifaceted phenomenon and, as a result, scholars have come up with various, slightly different conceptualisations.
Yet most of them include the two elements of conflict and salience (see e.g. De Wilde 2011;De Wilde et al. 2015;Statham and Trenz 2013). Clearly, some conflict surrounding an issue is required; without diverging attitudes there is no politicisation. In the case of European integration, this condition is met and, although the level of controversy varies across countries, it is generally substantial (Hoeglinger 2015: 128-30). Salience, the second essential element, is more demanding. Particularly during election campaigns, when a large number of political issues compete for visibility, public attention is scarce. Yet without visibility, citizens will fail to notice an issue no matter how bitterly contested it is by some marginal group, and hence there is no mass politicisation. For this reason, many scholars studying the politicisation of Europe have exclusively focused on salience (e.g. Green-Pedersen 2012; Kriesi 2007), and while politicisation is more than salience alone, salience is certainly its most critical aspect. 3 The closely related concept of issue-emphasis is based on the idea that the key to winning an election is for a party to succeed in putting its preferred issues on top of the political agenda. Parties deliberately choose to emphasise those issues that are favourable to them and de-emphasise unfavourable ones (e.g. Budge and Farlie 1983;Budge et al. 2001;Petrocik 1996;Riker 1986). Hence, if politicians think they will benefit from an increased politicisation of Europe, they will try to move Europe further up the political agenda.
Yet parties face a critical constraint when (de-)emphasising Europe. As Steenbergen and Scott (2004: 187-8) put it: 'Parties cannot decide on issue salience willy-nilly. Lest they find themselves marginalized from the mainstream national political debate, they will have to consider the importance that other parties attach to the issue of European integration' . This argument is shared by agenda-setting scholars who criticise traditional issue-emphasis approaches for being unable to explain the considerable issue overlap between individual parties regularly observed in empirical studies. After all, if parties only emphasised their own issues we would expect them to talk past each other most of the time. From an agenda-setting perspective, however, party interaction is to be expected even when that means parties have to talk about issues that are disadvantageous to them (Green-Pedersen and Mortensen 2010: 261). Politicians are 'riding the wave' of prominent issues because they want to appear responsive to public concerns (Sides 2006). Hence, we should expect the overall salience of European integration to influence how strongly parties emphasise this issue.

H3: The higher European integration is on the political agenda during an election campaign, the more strongly it is emphasised by each of the competing parties.
To conclude this section and bring together the various expectations from the literature outlined above, one can distinguish four scenarios of how Europe is being politicised (Figure 1). This allows us to schematically classify the different scholarly views, which diverge in their expectations about the extent of the politicisation of Europe and the nature of the underlying conflict structure. If the ideological characteristics of the parties are expected to primarily determine whether they politicise Europe, scenarios I and III are potential outcomes. Culturally conservative parties close to the TAN pole are probably those most strongly mobilising on this issue, but alternative ideological patterns of politicisation are also compatible with this view. Most of these scholars take a high level of politicisation for granted, assuming that the successful integration of Europe into larger and salient patterns of political conflict should greatly boost politicisation. Scenario I is therefore most in line with this literature. By contrast, if one believes that European integration issue-emphasis is mostly instrumental and unrelated to the ideological characteristics of the parties, there are no clear expectations as regards the extent of politicisation. Europe might be such a powerful issue that it is able to spur massive politicisation single-handedly (scenario II). Alternatively, its supposed strategic use primarily by opposition and fringe parties might be the very reason that the politicisation of Europe remains limited, as those that ultimately matter -the mainstream parties -lack the incentive to politicise it (Green-Pedersen 2012; scenario IV). In summary, there are good arguments in the literature for scenarios I, II and IV; only scenario III finds little backing. In the empirical section below, we will see which of these scenarios comes closest to reality. Of course, hybrid forms are a possibility -for example, both ideological and strategic motives might drive politicisation to varying extents.

Methods, data and operationalisation
In today's 'audience democracies' (Manin 1997), politics takes place primarily in the mass media. Citizens' perception of political issues, actors and political conflict in general are largely shaped by their representation in the media since this is the first place they turn to for information about politics. Consequently, the strategies of parties and candidates have adapted to this and they try to get their statements in the news to shape the public debate to their advantage. Against this backdrop, this article relies on a large-scale quantitative analysis of party statements in newspaper articles during election campaigns. Table 1 reports the newspapers as well as the elections and the number of statements that were coded for each of the countries under study. 4 An election campaign is defined as the two-month period before Election Day. In the randomly selected articles, party statements about any political issue were coded, be it from an official party speaker or any party-affiliated individual, such as a member of the executive branch. For each statement, the name of the individual and their organisational and party affiliation, the political issue, and the adopted position were coded. Statements are categorised into 12 issue categories: Europe, welfare, economic liberalisation, budgetary rigour, cultural liberalism, culture, immigration, army, security, ecology, infrastructure, and institutional reform. 5 Tests for intercoder reliability showed acceptable results. 6 When trying to get their preferred message across, politicians not only have to consider their competitors, but also the fact that the media has increasingly become an autonomous actor that follows its own logic, such as news values to which they need to adapt (Esser 2013). However, recent research argues that unlike during routine times, the media's agenda-setting role is relatively limited during electoral campaigns (van Aelst and de Swert 2009;Walgrave and van Aelst 2006: 96-8). First, during election campaigns, parties and candidates multiply their communication efforts and flood the media with their press releases and (pseudo) events to shape the public agenda in their favour. Although most  1975, 1994, 1999, 2002France 1978France , 1988France , 1995France , 2002France , and 2007Germany 1976Germany , 1994Germany , 1998Germany , 2002Germany , and 2005uK 1974uK (Feb.), 1992uK , 1997uK , 2001uK , and 2005netherlands 1973, 1994, 1998, 2002, 2003switzerland 1975, 1991, 1995, 1999, 2003. no genuine tabloid exists in the netherlands so a tabloid-style newspaper with a wide circulation was used instead. journalists are fully aware of this strategy, reacting with cynicism and a horserace style of political reporting, politicians still largely succeed in getting their messages across. Second, this is not least because during election campaigns the media devotes considerably more attention than usual to politics. This lowers the barriers to access for politicians of all types and provides them with opportunities to speak out. Third, the journalistic norm of fair and balanced reporting is particularly well respected by the media during election campaigns, especially because parties and candidates are quick to make allegations about unfair treatment by the media. Overall, this suggests that media coverage during election campaigns is less prone to potential bias. The salience of a political issue is operationalised as the share of statements dealing with this particular issue out of all topical statements made by the parties in an election campaign. Correspondingly, the second dependent variable, issue-emphasis, is the share of statements about the issue in question in relation to all issue statements made by a particular party.
The independent variables are operationalised as follows: For ideology, the location of the particular party on the economic axis and the cultural axis of the political space is used. The economic axis spans from the left pole to the right pole, the cultural axis from the TAN pole to the GAL pole. A party's score on these two axes is the salience-weighted index of its positions towards economic and cultural issues, respectively. The issues used for calculating the economic left-right axis score are welfare (-), economic liberalisation (+), and budgetary rigour (+). The issues of the cultural TAN-GAL axis are cultural liberalism (+), culture (+), immigration (-), army (-), security (-), and ecology (+). 7 Both axis scores range from -1 to +1. For example, a right-wing populist party with only culturally conservative (TAN) issue positions would receive the value -1.0 for the cultural axis, while a green party with only culturally liberal (GAL) issue positions would score +1.0. In reality, of course, most parties have scores in between these two extremes. Ideological extremism is operationalised as the squared Euclidian distance of a party from the centre location in the political space, i.e. the sum of the squared scores on the economic and cultural axis. Values can range from 0 to 2. The opposition party variable is operationalised as the share of time a party was in opposition during the 10 years before the election; it ranges from 0 to 1. Finally, the regressions control for systemic salience, the salience of European integration in the particular election campaign as a whole. In order to avoid endogeneity and following Steenbergen and Scott (2004), the systemic salience was calculated for each party individually by excluding the party's own statements about Europe.

Findings
Different theoretical strands hold competing expectations as regards the general extent of the politicisation of Europe and, related to this, they also disagree over the crucial explanatory factors that drive European integration issue-emphasis. In this empirical section, both questions will be assessed in turn. Figure 2 reports the salience of European integration in election campaigns across all six countries both over time and compared to other political issues. As the findings show, European integration seriously entered the political stage in the aftermath of the signing of the Maastricht Treaty in the early 1990s: Negligible in the 1970s with on average just 2.2 per cent of all public statements in election campaigns, the salience of Europe has grown considerably in the 1990s, peaking at 7.4 per cent on average in the second half of that decade. Subsequently, the salience of Europe decreased slightly but remained substantial throughout the 2000s, reaching 4.6 per cent in the latest period. Note that the bulk of the campaign discourse on Europe was about integration in general terms or about the big constitutive issues (polity dimension), whereas specific policies were only rarely discussed. 8 Benchmarked against other issues on the political agenda, the salience of Europe can be put into a broader perspective. The values for Europe are fairly similar to those for immigration, with the important difference that immigration exhibited a continuing upward trend in the 2000s, reaching its highest salience of 9.0 per cent in the latest period. The salience of both of these challenger issues, however, is still considerably lower than those of traditional key issues, as a comparison with the welfare issue suggests. On average, between 15 and 20 per cent of all party statements deal with welfare, making it the most salient issue on the political agenda. Moreover, if we look at the median ranking among the 12 political issues in terms of salience, the modest importance of European integration on the political agenda becomes even more apparent. Europe starts from a lowly 10th place in the 1970s and then climbs to the rank of 6.5 in the second half of the 1990s. But then again, Europe only comes in at the back of the  Figure 2. the limited salience of europe in domestic election campaigns -share of statements and ranking on the political agenda of european integration and selected issues.
field of relevant issues in the 2000s (rank 9 and 8.25, respectively). Immigration, by contrast, has become an issue to be reckoned with -with a median rank of 6.5 and 5 in the 2000s. Hence, when compared with other political issuesrather than with its own marginal salience back in the long-gone days of the permissive consensus -European integration is moderately politicised at best. This finding seriously questions the predominant scholarly claim of a rampant politicisation of Europe.
In the remainder of this section, the focus of analysis shifts from the entire system to the actor level. The results of a series of pooled OLS regressions testing the factors hypothesised to drive European integration issue-emphasis are reported in Table 2. Model 1 controls for systemic salience only, while the next three models separately test for ideology, extremism, and opposition-party status. Model 5 is the full model which includes all variables simultaneously. Figure 3 provides the marginal effects of Model 5 to facilitate interpretation (and comparisons between the issues of Europe and immigration, see below). Note that a dummy and interactions for the decade are included in all models The regression results demonstrate first and foremost that ideology has a systematic and strong effect on European integration issue-emphasis, as hypothesised. The effect of the cultural axis is most pronounced: The closer a party is to the culturally conservative TAN pole, the more it tends to emphasise European integration in election campaigns. Strongest in the 2000s, this effect is also visible in the 1990s, although the coefficient here is slightly smaller and only marginally statistically significant at the 90 per cent level. 10 There is also a negative effect of the economic axis, albeit weaker and observable only for the 2000s (in the full model). Taken together, these findings suggest that those parties with a proper ideological package -either culturally conservative or, to a lesser extent, to the economic left -tend to emphasise European integration more than their competitors.
In contrast, the findings lend no support to the relevance of purely strategic motives. A party in opposition, it was hypothesised, will put pressure on the government by emphasising Europe, while the same party tones down its anti-European rhetoric if it is in government. Similarly, extreme parties at both ends of the political spectrum might politicise Europe to challenge the mainstream parties. However, both of these expectations fail to manifest in the empirical analysis. To corroborate this finding, I tested alternative operationalisations of the originally interval-scaled opposition variable. Yet neither a dummy for being in opposition on the day of the election nor a trichotomous version performed any better. This suggests that parties do not politicise Europe systematically merely out of strategic considerations related to party competition when they are in opposition or ideologically extreme. This is not to deny that  Figure 3. the impact of the cultural axis, the economic axis, extremism and oppositionparty status on european integration and immigration issue-emphasis (marginal effects, with 95% confidence intervals).
many fringe parties mobilise against European integration. Yet they are not the only ones, and their anti-European attitude is more than a strategic choice; it is a substantively motivated attitude, as demonstrated by the results above. Moreover, there is a strong and statistically significant effect of systemic salience that persists throughout all the models. In line with the expectations of the agenda-setting literature and previous findings (Netjes and Binnema 2007;Steenbergen and Scott 2004), this supports the argument that parties are severely constrained in their individual issue-emphasis strategies by the overall agenda of the party system. Exploring what shapes this general agenda is beyond the scope of this article. Yet the finding highlights that it is essential to clearly distinguish between a systemic and an actor-specific component.
To check the robustness of the findings, I re-estimated the analysis in a number of different ways (see the online appendix for the detailed results). I ran the analysis on restricted sub-samples: without influential observations, without fringe parties (radical left and populist radical right), and without the UK and Switzerland (which some scholars consider as outlier cases in terms of attitudes towards Europe -e.g. Green-Pedersen 2010). The substantive findings generally hold across these sub-samples. In particular, the effect of the cultural axis is not driven exclusively by parties at the fringes. The pattern can also be found among mainstream parties (as the robust finding for the sub-sample without the fringe parties demonstrates). Hence, it is worth emphasising that not only the populist radical right but also moderate conservative parties are more likely to mobilise on Europe than their culturally more liberal counterparts. By contrast, the effect of the economic axis strongly decreases and becomes statistically non-significant when we remove the fringe parties from the sample. This suggests that the effect is indeed merely caused by opposition against Europe from the radical left. Moreover, I tested and performed various alternative estimation strategies. All of these results were very similar to the original analysis and, while the size of the relevant coefficients decreases in some of these estimations, they continue to be statistically significant and the substantive findings remain unchanged.
In the final step of the empirical analysis, I compare these patterns found for European integration with those for immigration. Both are newly emerging issues with which established parties struggle, whereas challenger parties are likely to benefit from their politicisation. Moreover, as outlined above, both issues are part and parcel of a new globalisation cleavage that pits the winners and losers of increasing economic, political and cultural denationalisation against each other (Kriesi et al. 2012). The far right-hand column in Table 2 reports the estimates of a regression similar to Model 5, yet with immigration issue-emphasis as the dependent variable. Figure 2 also shows the corresponding marginal effects. Interestingly, the direction of the effect of a party's position on the cultural axis on immigration issue-emphasis is the same as for European integration, but considerably larger -in both decades more than double the size than for European integration in the 2000s. Hence, parties at the conservative TAN pole of the cultural axis tend to emphasise immigration much more strongly compared to other parties than they do European integration.
The pattern for the economic left-right axis is also weaker and more ambiguous over time. Yet the direction of the statistically significant effect in the 2000s is exactly opposite to what we found for Europe: A party is more likely to emphasise immigration the closer it is to the economic right pole. These patterns for the two axes found for immigration correspond nicely with Kitschelt's (1995) 'winning formula' for right-wing populist success, which he argued is combining economic rightist and culturally conservative views. In fact, the patterns do not fit well with the globalisation cleavage argument, which would instead expect that cultural conservativism aligns with economic (leftist) opposition to globalisation. However, radical right parties have traditionally taken care to obscure their inconsistent views vis-à-vis globalisation by toning down their market-liberal, anti-redistributionist stance (Kitschelt 2007(Kitschelt : 1181 and by promoting welfare chauvinism instead.
Finally, unlike in the case of European integration, extremism has a positive and statistically significant effect on immigration issue-emphasis. Hence, radical parties from all sides, regardless of whether they are supportive or opposed, are more likely to emphasise immigration.

Conclusion
The present article set out to test differing expectations in the literature about both the crucial drivers of the politicisation of Europe and, related to this, the actual extent of this phenomenon. Although most observers are currently, due to the ongoing crisis, quick to jump to the conclusion that we have finally arrived in an age of rampant politicisation, this latter question is still open to debate. How do my findings compare to the four possible scenarios of politicisation laid out in Figure 1? With regard to the underlying conflict structure, the findings suggested that European integration issue-emphasis is indeed driven to a significant extent by ideological conflict and firmly aligned with long-standing, more basic political concerns. Parties close to the culturally conservative TAN pole, such as the populist radical right but also, remarkably, the mainstream conservative right, are more prone to emphasise Europe. There is a similar but considerably weaker pattern for parties close to the economic left-wing pole as well. Overall, this is well in line with theoretical arguments put forward by scholars such as Kriesi et al. (2006Kriesi et al. ( , 2008Kriesi et al. ( , 2012 and collaborators (2004, 2009) who stress the ideological component, and in particular the cultural dimension of contestation over Europe. By contrast, a party's opposition status failed to show any impact, and fringe parties at both ends of the ideological spectrum do not systematically put more emphasis on Europe. Evidence for either of these two effects would have lent support to the argument that the politicisation of Europe is a programmatically unanchored strategy to challenge mainstream and governing parties.
What about the second dimension that distinguished our scenarios, the magnitude of politicisation? Although European integration has gained considerably in prominence since the long-gone days of the silent permissive consensus, when compared to other political issues the salience of Europe remained at a relatively moderate level throughout the 1990s and the 2000s. While no longer fast asleep, the giant has also not fully awakened. Hence, we are left with scenario III, which most accurately reflects the general pattern of party politicisation of Europe in domestic election campaigns: a quite unexpected combination of a low to moderate level of politicisation, yet at the same time well-structured along pre-existing lines of political conflict. This does not fit well with any of the discussed theoretical strands: those who rightly view ideology as a critical driving factor generally tend to overstate the actual level of politicisation that results from it, while there is, on the other hand, no support for the competing claim that purely strategic considerations are essential for explaining the politicisation of Europe.
Before discussing the further implications of the findings, two caveats are in order. First, the relative impact of the parties' positions on the cultural GAL-TAN axis and the economic left-right axis, respectively, might have been different had we not selected six Western European countries but had instead chosen a Southern European sample. In these countries, the populist radical right is generally weaker and the radical left, by contrast, more prominent, giving economically motivated Euroscepticism a more powerful voice. Still, the general pattern of ideological embedment should be the same. Second, the findings of the present article do not apply to Eastern Europe, where the logic of the politicisation of Europe is different as the structure of party competition is radically dissimilar to that in Western Europe (see Marks et al. 2006).
How are we to make sense of the somewhat perplexing finding that the politicisation of Europe in domestic election campaigns is limited although the issue has become firmly ideologically embedded? First of all, the fact that a newly emerging issue is becoming aligned with traditional lines of political conflict seems not necessarily to lead to strong politicisation. In the case of European integration, the multiple and intricate linkages of this issue with the pre-existing political space may actually hamper politicisation (Hoeglinger 2015). Elite opposition to European integration is motivated by a diverse array of reasons, notably economic and cultural ones. Politicians face a thorny and often unresolvable dilemma as some of their ideological core concerns speak in favour of certain aspects of European integration, while others lead them to adopt a sceptical stance. As a result, parties have to cope with ambivalent attitudes as well as unclear, cross-cutting lines of conflict; the risk of strange bedfellows is looming. What is more, the meaning of what Europe is actually about is changing over time and is difficult for any individual actor to control in light of the highly diverse and contested framing (Helbling et al. 2010). Hence, the issue of European integration is thereby potentially undermining what parties generally strive for -a simple and coherent programmatic package based on a few core concerns. This makes politicisation unattractive.
The comparison of European integration with its twin issue of immigration was also instructive in this regard. As argued by Kriesi et al. (2012), both issues are part of an emerging new globalisation cleavage. However, there are notable differences between the two. Immigration is more salient, particularly in the 2000s, and parties close to the culturally conservative TAN pole emphasise immigration much more than European integration. Hence, not European integration, but immigration is the issue these parties first turn to in order to mobilise their constituencies. Moreover, immigration is emphasised most by parties with a culturally conservative and an economic rightist stance -a combination, unlike the one found for European integration, which corresponds nicely with Kitschelt's (2007Kitschelt's ( : 1995 'winning formula' for right-wing populist success. While both issues have become embedded in the political space, the conflict structure in the case of immigration is simpler and more clearly defined, and therefore the issue is easier to handle in political competition. Taken together, this strongly suggests that the new globalisation cleavage will primarily become manifest in the sustained politicisation of immigration, and only secondarily of European integration. One could also argue that the difference found between European integration and immigration is in line with the hard versus easy issue dichotomy by Carmines and Stimson (1980), which implies that issue voting is more widespread in the case of the latter issues (and, hence, politicising such an easy issue is electorally more rewarding).
However, recently things have become more complicated as some culturally conservative parties have increasingly focused on intra-EU migration by emphasising the alleged negative consequences of the free movement of persons for their 'own people' . They are mobilising electoral support with a topic that perfectly fuses these two hitherto relatively independent political issues. This closer coupling of the issues of European integration and immigration, and its consequences for their politicisation, certainly merit closer attention in future research. Notes 1. These labels were introduced by Hooghe et al. (2004). The cultural axis has also been labelled the new politics, postmaterialist (Inglehart and Flanagan 1987) or libertarian-authoritarian axis (Kitschelt 1994). 2. I call these motives instrumental and distinguish them from the ideological motives outlined above even though instrumental considerations can also play a role in the case of the latter. For example, a TAN party's emphasis on Europe can be conceived of not only as an ideologically motivated behaviour but at the same time also as a strategic choice. However, the crucial difference is that this choice takes place within highly restrictive ideological constraints, whereas the former strategy of picking any issue that is unfavourable to the established parties should, if actually applicable, be open to every fringe or opposition party regardless of their particular ideological outlook. 3. There are also some more technical advantages of focusing on issue salience and issue emphasis in order to study politicisation. In particular, the salience of an issue is quantifiable and its measurement quite straightforward, which is essential for a systematic comparison. 4. The dataset used was assembled within the framework of a larger project (Kriesi et al. 2012; National political change in a globalising world. Supply-side data on national election campaigns. University of Zurich and LMU Munich). 5. The online appendix provides a detailed description of the 12 issue categories into which the inductively coded topics were finally aggregated. The few miscellaneous topical statements that could not be assigned to any of the 12 issues (less than 2 per cent) were excluded from the analysis. 6. More details on the nuclear-sentence coding approach, including intercoder reliability, can be found in the online appendix. Hoeglinger (2015: Chapter 3) and Helbling and Tresch (2011) discuss in depth the advantages and limitations of this data source for political analysis and compare it with more traditional data. 7. The signs in brackets indicate how the issues are related to the corresponding axis. 8. A breakdown of the salience of Europe in election campaigns by individual countries, elections and by the different types of media outlets is provided in the online appendix. Also see the appendix for more detailed results on the internal composition of the European integration issue. 9. As potential country effects are already part of the systemic salience, country dummies need not be included. However, see the online appendix for an alternative estimation with country dummies. 10. Moreover, the cultural axis becomes non-significant for the 1990s in some of the alternative estimations (see below).