Microsoft Word - EV_185_Amadio_Angelucci.docx Italian Political Science, VOLUME X ISSUE X * Both authors contributed equally to this paper. Published in 2023 under Creative Commons BY-NC-ND license by: Italian Political Science. ISSN 2420-8434. All copyrights of article contents are retained by the authors. Volume X, Issue X, x–x. Contact Author: Davide Angelucci, LUISS Guido Carli University, Rome E-mail address: dangelucci@luiss.it Disconnected? Public Opinion, Economic Elites, and Political Parties during the Migration Crisis Maria Giulia Amadio Viceré* EUROPEAN UNIVERSITY INSTITUTE, FLORENCE Davide Angelucci* LUISS GUIDO CARLI UNIVERSITY, ROME Abstract This article analyses whether European political parties were responsive to the policy preferences of citizens and economic elites over immigration during the migration crisis. To do so, it derives hypotheses from the scholarly literature on party responsiveness and tests them on survey data collected in 2016 and 2017 from among voters, political parties, and economic elites in 10 different EU member states. Contrary to the widespread belief about the crisis of contemporary representative democracies, the article shows that political parties’ positions on immigration changed consistently with changes in public opinion. On the contrary, the article finds no significant relation between the positions of economic elites and those of political parties. These empirical results are par- ticularly relevant for the study of democratic representation, as they challenge the widespread assumption about the crisis of contemporary representative democracies. 1. Introduction his article addresses the state of democratic representation in Europe with specific regard to political parties’ representation of public opinion positions during the migration crisis.1 Democratic representation is defined as the ability of political elites to aggregate and represent, within democratic institutions, interests that are dif- fused in societies (Dahl, 1973). This function is mainly carried out by political parties, which traditionally aggregate societal interests and formally express them through elected representatives in national parliaments. Nonetheless, globalization and its impli- cations for domestic decision-making have catalysed the tension between responsible and responsive political parties (Mair, 2009; 2013). Not only are states now embedded in complex networks of interdependence at the international level, but the strengthening of ties with other states and international actors prompts political parties to behave respon- sibly. Against this backdrop, to respect countries’ external commitments, political parties 1 The article conceptualizes the migration crisis in a broad manner. On one hand, it considers both refu- gees and migrants arriving in Europe in the period considered. On the other, it conceives this crisis period as the ensemble of the years during which there was a significant increase in the influx of migrants and refugees to Europe (i.e., 2014-2017). T Disconnected? 2 have often gone against the preferences of their domestic electorates (Bardi et al., 2014, p. 237). By increasingly interpreting their roles as responsible elites, it has been argued, po- litical parties (particularly mainstream ones) have favoured the alienation of voters. The latter, in turn, have grown increasingly frustrated by parties’ lack of responsiveness to their policy preferences (Brady et al., 2019). Existing scholarship claims that such a phenomenon is particularly relevant in the European Union (EU). It is a widespread opinion that during the multiple crises that oc- curred within and around Europe in the 2010s, the EU underwent a generalized crisis of political representation (Kaldor and Selchow, 2013). While at the domestic level this cri- sis led to a ‘strain of representation’ (Rohrschneider and Whitefield, 2012), at the EU level it raised concerns about the democratic accountability of the EU responses to the challenges it had to face (Fabbrini, 2014). Indeed, the widespread criticism that current political parties are disconnected from their voters (Rasmussen, 2019) undermined the argument according to which, since national governments represent citizens’ prefer- ences, member states, by controlling the EU decision-making system, inherently make its decisions accountable (Moravcsik, 2002). Certainly, political parties’ responsiveness is a well-known issue in the scholarly literature (for a review see Beyer and Hanni, 2018). Nonetheless, this issue has only rarely been examined in the context of the recent crises the EU has faced. To address this research gap, the article examines one of the recent EU crises, namely the migration crisis. In particular, it aims to answer the following ques- tions: were political parties responsive? And, if so, to whom? To answer these questions, the article examines the association between the posi- tions of political parties and those of citizens on the issue of immigration from 2016 to 2017 in 10 different EU member states, namely: Czech Republic, France, Germany, Great Britain, Greece, Italy, Netherlands, Portugal, Poland, and Spain. On the one hand, the timeframe considered allows us to delve into the national divisions on immigration, and the ensuing formation of conflicting positions between member states over the managing of the migration crisis at EU level. Significantly, focusing on this time frame allows us to measure political parties’ political representation of voters on immigration during the implementation of the 2015 EU temporary relocation scheme (Council of the EU, 2015a; 2015b), and hence in a period when immigration returned to the centre of political debate across Europe. On the other hand, examining these member states in a comparative re- search design allows us to achieve stronger confirmation of our analytical claims. Our analysis is based on a combination of survey data collected by the EuEngage pro- ject (Cotta et al. 2021) among voters and economic elites, with data from the Chapel Hill Expert Survey (CHES) (Jolly et al. 2022) dataset on political parties’ policy positions. At the same time, our article considers not only the relation between public opinion and political parties, but also the one between political parties and economic elites. With few exceptions (e.g., Conti et al. 2021), the extant scholarly literature has largely neglected such a relationship. Given the relevance of this aspect, scholars’ lack of attention is surprising. As a matter of fact, economic elites may express positions on policy issues that differ from those of public opinion. Overall, our results bring into question the widespread assumption that contempo- rary representative democracies are in crisis. In fact, the empirical analysis demonstrates that political parties aligned their preferences over immigration with those of their AMADIO VICERÉ and ANGELUCCI 3 citizens, rather than with those of their national economic elites. While doing so, political parties adopted increasingly restrictive policy positions towards migrations flows. Nota- bly, the analysis shows that their preferences in this regard were associated with the public’s perception of immigration as a threat. The remainder of the article is structured as follows. First, we place the migration crisis within the context of the scholarly debate on the crisis of democratic representation and discuss its relevance for our analysis. Second, we outline our analytical framework. Third, we provide an overview of our research design, our data sources, and research strat- egy. Fourth, we present and discuss the empirical results of the analysis carried out. Fifth, and finally, we draw conclusions from our study. 2. Policy Representation and the Migration Crisis Has the representative capacity of parties decreased? And, if so, why? These are key ques- tions in the scholarly debate on democratic representation and, more generally, on the state of representative democracy (Andeweg, 2019; Mair, 1995; 2011; Schmitter, 2001). Political parties can secure democratic representation in different ways. Nonetheless, the existing scholarship on representation and on the quality of representative democracy claims that the tension between responsible and responsive forms of representation has thrown contemporary representative democracies into crisis (Mair, 2009; Lefkofridi and Nezi, 2020). Indeed, the distinction between responsive and responsible elites neces- sarily implies two different understandings of representation. Responsiveness consists of the ability of political parties to collect and represent cit- izens’ policy preferences consistently. It refers to the connection between public opinion preferences for policy change (or changes in opinions) and changes in policies (or changes in parties’ policy positions) (Beyer and Hanni, 2018). As such, responsiveness is a relevant goal for democratic governments (Erikson, 2013). Responsibility, in turn, re- fers to the willingness of representatives to abide with accepted norms, procedural rules, and conventions, such as the ones set by international treaties and institutions (Lefkofridi and Nezi, 2020, p. 335). By acting as responsive elites, representatives gener- ally follow the short-term orientations and preferences of their constituencies, be they common citizens or economic elites. In these cases, the representation of societal inter- ests within democratic decision-making institutions is substantially based on the reflections of such preferences and their variations. By acting as responsible elites, in turn, representatives carry out their activities based on a relationship of trust with voters, epitomized by their electoral mandate. Even though voters subsequently evaluate their performances, in these situations political parties act as interpreters of society’s general interest without being directly constrained by the voters’ preferences (Mair, 2009). The migration crisis can be considered as the epitome of what has been called the responsive versus responsibility dilemma (Lefkofridi and Nezi, 2020). Although to dif- ferent extents, most EU member states experienced unexpected waves of migration because of the difficult political transition occurring in the Middle Eastern and Northern African region following the 2010-2011 Arab uprisings. While EU responses to the migra- tion crisis were characterised by inefficiencies and stalemates, the pre-eminence of intergovernmental practices in such responses fuelled mistrust between member states, as well as between some national governments and the EU institutions (Amadio Viceré, Disconnected? 4 2019). As the migratory pressure on the European continent augmented, the constraints imposed by international law as well as by EU policy frameworks reduced the decision- making space for domestic political parties (Moravcsik and Schimmelfennig, 2019). Meanwhile, political parties had to deal with a public opinion that was less and less keen to accept such rules. In fact, as the migratory flows towards Europe augmented, public opinion throughout the continent increasingly supported restrictive policies against im- migration (Debomy and Tripier, 2017). Crucially, such policy preferences were often in contradiction with the constraints imposed by international law as well as by agreements within the EU institutional framework. Against this backdrop, immigration became highly politicized across Europe, turn- ing into a key issue of political competition in many countries (Grande et al., 2019). It became a winning issue especially for radical-right parties, as well as for those parties which were previously on the edge of political competition. By spearheading voters oppos- ing immigration, the latter managed to reshape the political landscape of many European countries. Eventually, the success of radical-right parties in this regard also pushed main- stream parties, which generally adopt less critical positions on immigration, to gradually adopt tougher positions towards migration flows (Abou-Chadi and Krause, 2018; Urso, 2018; Van Spanje, 2010). All this notwithstanding, the relationship between the policy preferences of political parties and the orientation of public opinion on immigration has received scant attention so far. While existing studies on these matters focus on inter-party relations, they over- look the demand side of politics. In other words, they neglect the role of public opinion and other societal actors in the formation of political parties’ policy preferences regarding immigration. Certainly, there are studies that examine the link between public opinion and political parties on immigration (Angelucci et al., 2020), but they adopt a cross-sec- tional and static design (i.e., they focus on congruence). By doing so, such studies neglect the evolution of the relationship between the demand side of politics (i.e., demands com- ing from the bottom) and the supply side (i.e., responses from the top). Hence, they do not consider the dynamic dimension of political parties’ responsiveness. Of course, we are aware that in practice the connection between voters and parties may not always unfold in the same direction. For instance, studies on public opinion have demonstrated that parties may not only respond to voters’ positions but also shape their views (Steenbergen et al., 2007), including on immigration (Vrânceanu and Lachat, 2021). While acknowledging the complexities of the relationship between parties and vot- ers, in this article we specifically focus on the association between citizens’ and economic elites’ positions with those of political parties. 3. The Analytical Framework 3.1. Responsive to whom? Ever since the publication of Dahl’s seminal work on polyarchy (1971), the capacity of pol- iticians and political parties to respond consistently to the demands of public opinion is considered a key feature of representative democracies (Thomassen, 1994; Katz, 1997; Powell, 2000). It should therefore come as no surprise that the scientific scholarship fo- cuses extensively on the study of representation and, more specifically, on the link AMADIO VICERÉ and ANGELUCCI 5 between political parties and voters in terms of their policy preferences and issue priori- ties. In this context, scholars have generally either focused on congruence or on responsiveness. While both congruence and responsiveness can be considered forms of representation (Lax and Phillips, 2012), they hint at two distinct phenomena. Congru- ence is a static measurement, insofar as it relates to the correspondence of mass public preferences with those of political elites at a given point in time. It has been traditionally employed to determine the left-right ideological proximity between voters and political elites (Miller and Stokes, 1963). Comparisons between the ideological positions of mass publics and political elites at one point in time, have recurrently shown fairly robust levels of congruence. Such comparisons indicate that voters and parties would substantially align themselves along the left-right scale (Miller et al., 1999). When it comes to congru- ence on specific issues, however, results tend to be much more uncertain. Angelucci et al. (2020) show, for instance, that there has been a consistent gap between voters’ and par- ties’ positions on the process of European integration, with elites being systematically more Europhile than citizens in different policy areas. Similarly, by examining the rela- tionship between elites and voters on immigration policy, Brady et al. (2020) provides clear evidence of wide gaps between voters and parties. Responsiveness, instead, is a dynamic concept. Rather than focusing on the static correspondences between political parties’ positions and those of voters, it looks at the ‘correlation between prior public opinion and policy outputs’. At the same time, it exam- ines whether ‘changes in public opinion lead changes in policy in the same direction’ (or changes in policy positions of representatives) (Beyer and Hanni, 2018, p.18). Based on this definition, we conceive responsiveness as being not only in terms of policy outputs, but also in terms of the capacity of parties to shift their policy positions consistently with changes occurring in public opinion. Examining the dynamic responsiveness of political parties, therefore, necessarily re- quires a focus on their adaptation to the stimuli coming from voters, as well as from other political and societal actors, such as economic elites and interest groups. Looking at the relationship between voters and parties, assuming the rationality of parties, one would expect the latter to adapt their positions to shifts in the policy positions of their voters (Stimson et al., 1995). Indeed, acting in line with voters’ preferences can increase political parties’ chances of re-election. The underlying mechanism would be a punishment-re- ward one: political parties would respond to public opinion to avoid the potential negative consequences stemming from their lack of responsiveness at the elections. In this re- spect, empirical evidence has reported high levels of responsiveness both in the US and in Europe (Page and Shapiro, 1983; Stimson, 1991; Stimson et al., 1995; Brettschneider, 1996; Hakhverdian, 2012). However, in addition to citizens, other political actors may influence political par- ties’ policy positions (Klüver and Pickup, 2019). A growing body of scientific research has shown that there is a significant variation in the degree of responsiveness of political par- ties towards different subgroups or organised interests (Druckman and Jacobs, 2011; Gilens and Page, 2014; Jacobs and Page, 2005). For example, examining policy positions over the EU, Sanders and Toka (2012) found that political parties are more responsive to economic interest groups than to public opinion at large. Along similar lines, Klüver and Disconnected? 6 Pickup (2019) showed that the role of interest groups is decisive in structuring elites’ re- sponsiveness to public opinion on a number of issue areas. After all, this is not surprising. While it is reasonable to expect political parties to be rationally responsive to public opinion for fear of electoral consequences, it is also reason- able to expect political parties to respond to economic elites and interest groups. Significantly, political parties may respond to such groups even if they have positions which differ from those of the party’s electorate. In fact, in exchange for political influ- ence, interest groups might provide economic resources to support party campaigns, electoral support, and information when elections are being held (Giger and Klüver, 2016; Hall and Deardorff, 2006; Klüver and Pickup, 2019). Bearing all this in mind, it is reasonable to assume that, depending on the incentives in place, political parties may ra- tionally respond either to their voters or to economic elites. In this article, we posit that the characteristics of the issue at stake determine to whom political parties choose to be responsive. In particular, we argue that party respon- siveness to voters is conditioned by the extent to which an issue is contentious and/or salient in the public debate. According to Costello et al. (2020), when an issue is highly salient, parties have the incentive to fill potential issue gaps between themselves and their electorates. As a matter of fact, salient issues are highly visible in public debates and easily recognizable by voters, who are likely to make their electoral choices based on parties’ of- fers on these issues (or lack thereof). In other words, when issues are key for political competitions, the punishment-reward mechanism is likely to be activated. On the con- trary, when issues are out of the public opinion’s radar, parties are less likely to be evaluated on them. Consequently, they retain a wider room for manoeuvre. On these oc- casions, parties have greater possibilities to adopt policy positions which differ from their own electorate’s preferences and accommodate interests expressed by economic elites. It is relevant to point out that this argument resonates with the liberal intergovern- mentalist understanding of the formation of member states’ preferences in the EU context (Moravcsik, 1997). According to this theory, political elite goals are expected to vary ac- cording to changing pressure from national economic elites. Powerful interest groups may even capture government institutions and use them for pursuing their specific objectives, hence determining a bias of the political representation in their favour (Moravcsik, 1997). Nonetheless, when policy issues have significant implications for EU citizens’ lives, they generate more diffused patterns of societal interests (Hix, 2018). On these occasions, na- tional governments are likely to be tightly constrained by mass publics. As far as immigration is concerned, it is relevant to stress that this issue has become increasingly salient over the last decade. Immigration has in fact turned into a source of political contestation both in the EU and within member states (Van der Brug et al., 2015), to the point that it is currently considered the most prominent electoral issue of recent times (Green-Pedersen and Otjes, 2019). Considering all of this, we expect po- litical parties to be more responsive to common citizens’ positions on immigration than to those of economic elites. 3.2. Responsive to what? At this stage, our analytical framework provides some useful assumptions about the for- mation of political parties’ preferences. It predicts which societal interests – be they the AMADIO VICERÉ and ANGELUCCI 7 electorate’s interests or those of economic elites – are more likely to be represented under certain circumstances. However, our framework does not explain the stimuli to which po- litical parties are likely to respond. In other words, it does not tell us what changes are necessary, in public opinion positions and those of the economic elites, for the positions of parties to vary on a given issue. In this article, we posit that public opinion’s perceptions about immigration might have an impact on the variations of policy preferences about immigration among political parties. We base this argument on two different strands of literature. First, there are stud- ies suggesting that public opinion policy preferences about immigration are, among other factors, dependent on the perception of threat related to migration flows (Basile and Ol- mastroni, 2020; McLaren, 2003; McLaren and Johnson, 2007). In essence, the more threatening migration flows are perceived to be, the more favourable are likely to be the public opinion positions towards more restrictive policies on immigration. Second, a well-established literature on the effects of public opinion orientations on political elites’ foreign policy preferences shows that such preferences become more supportive of tougher and restrictive policies towards an external enemy or danger, if public opinion perceives such an enemy or danger to be particularly threatening (Cohen, 1979; Holsti, 1996; Hurwitz and Peffley, 1990; Peffley and Hurwitz, 1992). Combining the analytical argument about the public opinion’s perception of immi- gration as a threat and the argument about political parties’ responsiveness to public opinion’s threat perceptions, we can reconstruct the mechanism that urges political par- ties to be responsive to public opinion in this issue area. By conceptualizing immigration as an external threat, it is plausible to argue that the connection between political parties and public opinion on immigration can be found in the relationship between voters’ threat perception and the policy responses of the politicians. In other words, if political parties were responsive to public opinion positions on immigration, any shift in their pol- icy preferences would derive from changes in public opinion’s threat perception of immigration. All in all, considering the arguments outlined so far, we expect that: 1) political par- ties will be responsive to the preferences of public opinion on immigration rather than to those of economic elites; and 2) political parties will adopt increasingly restrictive policy positions on immigration in response to increases in the public opinion’s perception of immigration as a threat. 4. Data and methods 4.1. The research design The literature on policy congruence has often assessed the quality of representation by comparing the positions of political parties with those of public opinion on the same issue. This assessment is typically conducted through a correlation analysis or through a com- parison of the average positions of politicians and public opinion on given issues.2 The rationale behind these empirical methods is simple: the greater the correlation, the greater the level of congruence. However, it is important to note that these empirical 2 For a discussion of different strategies to measure congruence see Wlezien (2016). Disconnected? 8 analyses are generally carried out at a single point in time. As a corollary, they do not fully succeed in evaluating the elites’ actual level of response to changes in the positions of pub- lic opinion (Simonovits et al., 2019). The changes occurring over time in the positions of political parties in relation to changes in the positions of the public masses (i.e., responsiveness) are relatively less ex- plored. This may be because there are technical constraints deriving from the operational complexities of studying representation. The main problem in analyses of the relation- ship between masses and parties is that they are observations belonging to groups that are distinct and separate. To overcome this problem, these groups can be compared at a macro level, measuring the positions of both public opinion and parties at the country level. An alternative solution is to create a bridge for comparability by measuring the po- sitions of political elites, public opinion, and economic elites at a meso level, namely that of political parties. Such a bridge would allow a comparison carried out in terms of party electorates and party positions. At first sight, adopting such an approach would leave two questions open, namely: who should the parties respond to? Would they respond to their constituencies or to voters at large? Crucially, however, positive answers to such ques- tions are not mutually exclusive. On the contrary, recent studies clearly show how political parties act to maximize support outside their own electoral bases without jeop- ardizing their traditional electoral bases (De Sio and Weber, 2014). Bearing all this in mind, in this article we decided to focus on the meso level of polit- ical parties. Therefore, we will first map the evolution over time of the perception of immigration as a threat by different parties’ electorates, both for public opinion and eco- nomic elites. Then, we will match these variations with the changes in the policy positions of the respective political parties. This approach has a clear advantage over a macro-level approach. Of course, the latter remains relevant from a descriptive point of view. None- theless, a party-level approach allows us to expand the number of observations in our analysis, and hence to have a more robust assessment of party responsiveness. In essence, rather than focusing on the overall differences at country level, we devote our attention to the multiplicity of parties that compete in the existing party system in each of the coun- tries considered. 4.2. The data Our analyses are based on two data sources: the EuEngage mass and economic elite panel (2016 and 2017) (Cotta et al. 2021), and the Chapel Hill Expert Survey (CHES), 2014 and 2017 (Jolly et al. 2022). The EuEngage dataset includes 11,638 respondents (mass sample) and 1,271 businesspeople (economic elite sample) in 10 European countries, who were in- terviewed for the first time in 2016 and subsequently re-interviewed in 2017.3 After removing the speeders,4 the mass panel consists of 10,928 observations, while the 3 It is relevant to note that, in addition to the Mass and Businesspeople surveys, the EuEngage dataset in- cludes an Elite survey which was conducted among politicians in the same 10 EU member states in 2016 and 2017. Elite surveys and expert surveys are often used to estimate party positions (Ecker et al. 2022). Still, we chose to use the CHES expert survey rather than the Elite survey to maximize the number of ob- servations. In fact, since elite surveys suffer from low response rates, they are often marred by a limited capacity to capture the positions of a large number of parties. 4 In particular, we removed those people answering in less than 50% of median time in each country in (at least) one of the two waves. AMADIO VICERÉ and ANGELUCCI 9 economic elite sample consists of 1,203 observations. This dataset is an optimal source for testing our hypotheses for a series of reasons. First, the time frame of the survey is particularly interesting, given that the two waves were collected during the implementation of the 2015 EU temporary emergency relocation scheme, which set mandatory relocating obligations for EU member states. Crucially, during this period the migration crisis returned at the centre of the political de- bate, both in the EU and domestic level. Second, the dataset’s panel structure allows us to map real changes occurring within the public and businesspeople samples about immi- gration. Finally, the dataset provides the opportunity to test our hypotheses in 10 European countries. Not only do these countries represent different geographical areas, but they also differ considerably in terms of history, tradition, political culture, economy, and society. In particular, earlier studies have demonstrated that, in these countries, the public opinion and political elites’ preferences regarding immigration were determined by: their different contextual factors (i.e., level of exposure to the crisis), their different predispositions towards migrants and the EU, and by their identity and ideological orien- tations (Basile and Olmastroni, 2020). Furthermore, while in some of these countries there were elections between the first and the second wave (i.e., Czech Republic, France, Germany, Netherlands, Spain), in the others there were no elections in the period consid- ered. This variance will not only make our findings more robust, but it will also allow us to decrease the possibility of detecting changes of perceptions in the electorates generated by the specific electoral dynamics of individual countries. The CHES dataset, in turn, estimates the parties’ positions on a wide range of issues, including policy preferences over immigration in more than 30 countries. These coun- tries include the 10 countries covered by the EuEngage dataset. The 2014-2017 datasets, especially, allow us to specifically analyse changes in policy preferences that happened in this period. In our understanding, responsiveness derives from the consistency between change in the position of voters and change in the position of parties on immigration. Put simply, if political parties changed their positions on immigration consistently with changes on immigration among their electorates, then they would be responsive to their constituencies. Therefore, our dependent variable consists of the change of party posi- tions on immigration, measured by the CHES data. Specifically, it is the difference between the position of the parties on immigration in 2017 and the position of the same parties in 2014. To measure the position of parties in the two waves, we considered a var- iable ranging from 0 (i.e., party fully opposed to a restrictive policy on immigration) to 10 (i.e., party fully in favour of a restrictive policy on immigration). We standardized both variables so that they could range from 0 to 1 in both waves. Then, we calculated the dif- ference between the two waves. This yielded a variable ranging from -1 (i.e., parties have changed towards a more open position towards immigration) to +1 (i.e., parties have adopted a tougher position in immigration). Our key independent variable is the change in the perceptions of immigration as a threat by public opinion and economic elites. Here, we rely on the EuEngage datasets of 2016 and of 2017 and, especially, on an item asking the respondents to express how threat- ened they felt by immigration. The variable ranges from 1 (i.e., do not feel threatened) to 4 (i.e., feel threatened). In each wave, we averaged respondents’ positions as sorted by the party they intended to vote for in general elections. We did so for public opinion and Disconnected? 10 businesspeople separately to measure the threat perceived by each of these two groups (again, by political party). In the two waves, the variables were standardized to range from 0 (i.e., do not feel threatened) to 1 (i.e., feel threatened). Then, we calculated the change in threat perception for each electorate as the difference between the perception of threat in 2017 and in 2016. Once again, this yielded a variable ranging from -1 (i.e., the electorate has become less threatened) to +1 (i.e., the electorate has become more threat- ened). After having calculated parties’ and electorates’ scores, we matched them in a unique dataset, consisting of an overall number of 57 party/electorate observations in 10 countries (i.e., the parties for which we have data in the mass survey, the businesspeople survey, and the CHES dataset). In addition to changes in threat perceptions among public opinion and economic elites, we also included as a series of other control variables: the electoral strength of the party, measured as the difference in votes share of the party in the two last general elec- tions available in the dataset, the left-right ideological position of the party,5 and its position in government/opposition. We constructed this last variable to account for the variation in government positions of parties between 2014 and 2017. In particular, we constructed a categorical variable which differentiates between: a) parties that were in government in 2014 and remained in government in 2017; b) parties that were in govern- ment in 2014, but then moved to opposition in 2017; c) parties that were in opposition in 2014 and then moved to government in 2017; and d) parties that were in opposition in 2014 and remained in opposition in 2017. Lastly, we included a dummy variable to distin- guish between countries where elections were held between 2016 and 2017 (i.e., the years when the two EuEngage waves were collected) and countries where there were no elec- tions in the same period. 5. Results According to our expectations, when it comes to highly politicized and contentious issues, political parties should be more likely to adopt policy positions in line with public opinion preferences. Given the highly politicized nature of immigration during the migration cri- sis, we expect political parties’ policy preferences to correspond to those of public opinion. To test these expectations, we set up a series of regression analyses to understand whether or not political parties were responsive to their voters on the immigration issue between 2016 and 2017. As discussed above, our dependent variable is the change in the policy preferences of political parties on immigration, while our key independent variables are the variations in perception of immigration as a threat by common citizens and businesspeople. Con- cerning the details of our empirical analysis, we estimated five linear regression models, including a random intercept to account for country variations. In each model, we let the intercept vary across country. By doing so, we were able to account not only for the vari- ance of the dependent variable that is explained at the individual level (i.e., electorate/party combinations), but also at the country level. Through this modelling 5 The left-right ideological position of political parties is calculated as the average score between left-right positions as measured in 2014 and 2017 in the CHES dataset (LRGEN item). In both cases, the left-right ideological position was measured on a 0-10 scale, where 0=Left and 10=Right. AMADIO VICERÉ and ANGELUCCI 11 strategy, we could estimate the effects of threat perception among citizens and business- people, net of contextual factors linked to cross-countries differences. The results of our analyses are reported in Table 1. In the first place, we assessed to what extent our dependent variable is explained either at country or individual level. To this end, we ran an empty model, without predictors (Model 1). The random parameters at the bottom of the table show that the variance of the dependent variable is mostly ex- plained at the individual level, that is to say at the level of the electorate-party combination. At the same time, the random parameters show that the variance at the country level is relatively small. This is reassuring. In fact, this result indicates that there is not much variation in the dependent variable due to contextual factors. Therefore, var- iations in party responsiveness may be found in the relation between parties and voters. The latter include, of course, both common citizens and businesspeople. Table 1. Random intercept linear regressions. Dependent variable: party positions’ change on immigra- tion. Model 1 Model 2 Model 3 Model 4 Model 5 Δ Immigration threat (Citizens) 0.260* 0.233* 0.366** (0.107) (0.111) (0.134) Δ Immigration threat 0.0951 0.0601 0.0511 (Businesspeople) (0.0641) (0.0647) (0.0658) Electoral strength of the party 0.0011 (0.00228) Position in government: Party always in opposition Baseline Party moved to government 0.0452 (0.0421) Party moved to opposition -0.0709 (0.0560) Party remained in government 0.0233 (0.0360) Left/Right position of party -0.0105 (0.00663) Election in 2016-2017 0.0211 (0.0376) Intercept 0.0222 0.0754* 0.0226 0.0702* 0.137* (0.0256) (0.0317) (0.0244) (0.0317) (0.0595) Random Parameters Country level variance 0.068 0.058 0.063 0.055 0.0393 (0.024) (0.022) (0.023) (0.022) (0.027) Individual level variance 0.102 0.099 0.101 0.099 0.095 (0.011) (0.010) (0.010) (0.010) (0.011) N 57 57 57 57 57 AIC -79.79 -83.29 -79.93 -82.13 -77.59 BIC -73.66 -75.11 -71.76 -71.92 -55.12 Notes: standard errors are reported in parentheses. As for the more substantial findings of our models, we assessed the association be- tween the threat perception of common citizens (Model 2) and businesspeople (Model 3) and variations in party positions across 2016 and 2017. These bivariate regression anal- yses confirmed our expectations. We found a significant and positive effect of threat Disconnected? 12 perception variations among common citizens on party position variations. Indeed, our results show that with any unitary increase in the threat perception of citizens, there was an increase in party preferences for tougher policy measures (b=0.260; p<0.05). We found the same positive effect for the relation between changes in the threat perceptions of businesspeople and changes in party positions. In this case as well, political parties seemed to react to the mood of economic elites consistently. Nevertheless, it is important to note that the coefficient of threat perception does not reach the conventional threshold of statistical significance. In Model 4, we included the two variables together (i.e., threat perceptions of both citizens and businesspeople). By doing so, we further confirmed the significant positive effect of threat perception of public opinion on political parties and the insignificant ef- fect of threat perception of businesspeople on political parties. These results provide some support for our expectations: when it comes to highly politicized issues, such as immigra- tion, political parties adapt their policy preferences to perceptions that are mostly diffused within the public opinion, rather than to those of economic elites. Finally, in Model 5, in addition to our two key independent variables, we plug into our regression analyses a set of control variables, namely: the electoral strength of the party, the position in government-opposition of the party, its ideological position (i.e., left-right ideological dimension), as well as a dummy variable distinguishing countries where elections were held between 2016 and 2017, and countries where there were no elections in the same pe- riod. The robustness and the strength of the effect of threat perception of common citizens on party position changes are not wiped out by the inclusion of these control var- iables. This result confirms once again the solidity of the relation between citizens’ threat perceptions and political parties’ reactions. Overall, this suggests that parties were actu- ally responsive to public opinion’s perceptions on immigration during the migration crisis. Finally, while we did not advance specific hypotheses on the relations between the other predictors we included as control variables and party position variations between 2016 and 2017, it is relevant to note that none of these variables reached conventional thresholds of statistical significance. The different status of the political parties consid- ered, namely whether they were in the government or in opposition, did not produce any effect on the changes of their positions. This result is somewhat surprising. Indeed, it would have been reasonable to expect that, compared to parties in opposition, governing parties would have had more constraints on changes in their positions, given their insti- tutional roles. At the same time, right-wing parties were not more likely to move towards tougher positions on immigration than left-wing parties. This result is in line, however, with the assumption that left-wing parties too have often adopted more restrictive posi- tions towards immigration as a reaction to the surge of radical right parties (Bale et al. 2010). Lastly, neither the electoral strength of parties nor the presence of an election pro- duced an effect on the variation of parties’ positions on immigration. 6. Discussion and Conclusions By examining the association between the positions of political parties, economic elites, and public opinion on immigration across Europe during a crucial period of the migration crisis, we were able to confirm the expectations stemming from the analytical framework AMADIO VICERÉ and ANGELUCCI 13 we elaborated. The article demonstrates that political parties aligned their policy prefer- ences on immigration, which is a highly salient issue, in accordance with citizens’ preferences rather than with economic elites’ preferences. In particular, our results sug- gest that parties reacted significantly to the stimuli provided by public opinion on the issue of immigration, shifting their policy positions consistently with changes occurring among citizens. Certainly, our results do not allow us to reach clear-cut and definitive conclusions about ‘who followed whom’ in the period considered. In this sense, our study should be considered as exploratory research on party-public opinion dynamics. As the scholarly literature shows, the connection between citizens and political parties may take different causal directions (e.g., Steenbergen et al., 2007): if it is true that parties’ positions can be influenced by public opinion, it is also true that parties, by cueing citizens, are in fact able to influence their opinions (e.g., Carsey and Layman, 2006; Lenz, 2012; Leeper and Slothuus, 2014; Slothuus and Bisgaard, 2021). Nonetheless, these results have important implications for those who study democratic representation. Previous studies demon- strated that public opinion and political parties do not differ much in terms of their policy positions on several issues, including immigration (Angelucci et al., 2020). In our article, we demonstrated that these positions were not only similar, but also that they changed consistently over time. Of course, the limited geographical and temporal span covered by this study does not allow us to simply generalise our findings. Nonetheless, this result is interesting in itself: if parties shift their positions consistently with citizens’ positions, is democratic representation in trouble, as many argue? Indeed, one may reasonably con- clude that our results disconfirm the widespread belief that contemporary representative democracies are experiencing a crisis; or rather that they indicate that the causes of such a crisis should be searched for elsewhere. Further research, for example, should investigate the patterns of political represen- tation on issues which are different from immigration. Indeed, one could also argue that the lack of connection between political parties’ positions and the economic elites’ posi- tions on this issue is due to the different relevance that this issue has for different societal groups. In this sense, immigration might not be salient enough for economic elites to push them to mobilize their resources to influence the policy preferences of political par- ties. Recent research in the US, analysing responsiveness on a broad set of issues, has produced different results compared to ours, showing that ‘the majority does not rule’, whereas powerful economic and business elites do have a strong influence on policy mak- ing (Gilens and Page, 2014, p. 576). These seemingly contradictory results suggest that the political representation of different societal groups might in fact be conditional on a num- ber of factors, including the nature and the salience of the issue at stake. From a more general point of view, to understand thoroughly the dynamics of political representation and identify ‘who rules’, future research will have to explore these factors further. Acknowledgements The authors would like to thank the LUISS Italian Centre for Electoral Studies (CISE), especially Professor Lorenzo De Sio, for organizing the seminar that inspired this article. Disconnected? 14 The authors would also like to thank the journal editors and the anonymous reviewers for their very useful comments. References Abou-Chadi, T. and Krause, W. (2018) The Causal Effect of Radical Right Success on Main- stream Parties’ Policy Positions: A Regression Discontinuity Approach. British Journal of Political Science 50(3): 829-847. Amadio Viceré, M. G. (2019) The International Context 2013-2018. In: J. Newell (ed.) The Ital- ian general elections of 2018. Italy in Uncharted Territory. Basingstoke: Palgrave Macmillan, pp. 73-94. Andeweg, R. B. (2019) Peter Mair on representative democracy. Irish Political Studies 34(2): 145-153. Angelucci, A., Smets, K. and Isernia, P. (2020) Representation at Work: Miller and Stokes go to Brussels. In: M. Cotta and P. Isernia (eds.) The EU through Multiple Crises. Represen- tation and Cohesion Dilemmas for a “sui generis” polity. Routledge: London, pp. 158-185. Bale, T., Green-Pedersen, C., Krouwel, A., Luther, K. R. and Sitter, N. (2010) If you can’t Beat them, Join them? Explaining Social Democratic Responses to the Challenge from the Populist Radical Right in Western Europe. Political Studies 58(3): 410–426. Basile, L. and Olmastroni, F. (2020) Sharing the burden in a free riders’ land: The EU migra- tion and asylum policy in the views of public opinion and politicians. European Journal of Political Research 59(3): 669-691. Beyer, D. and Hänni, M. (2018) Two Sides of the Same Coin? Congruence and Responsiveness as Representative Democracy's Currencies. Policy Studies Journal 46(S1): S13–S47. Brady, D., Ferejohn, J., and Paparo, A. (2020) ‘Are we losing touch?’ Mainstream parties' fail- ure to represent their voters on immigration and its electoral consequences. Italian Political Science Review/Rivista Italiana Di Scienza Politica 50(3): 398-421. Brettschneider, F. (1996) Public Opinion and Parliamentary Action: Responsiveness of the German Bundestag in Comparative Perspective. International Journal of Public Opinion Research 8 (3): 292– 311. Carsey, T. M. and Layman, G. C. (2006) Changing sides or changing minds? Party identifica- tion and policy preferences in the American electorate. American Journal of Political Science 50(2): 464– 477. Conti, N., Marino, B. and Memoli, V. (2021) Business Elites and European Integration: Really a Honeymoon Story? JCMS: Journal of Common Market Studies 59(6): 1400– 1418. Chen, R. (1979) Threat Perception in International Crisis. Madison: University of Wisconsin Press. Cotta, M., Isernia, P., Basile, L., Borri, R., Marangoni, F., Olmastroni, F. and Verzichelli, L. (2021) EUENGAGE Dataset 2016-2018 (v.2.0, 8 August 2021). Council of the EU (2015a) Council Decision (EU) 2015/1523. Brussels, 14 September, https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=OJ%3AJOL_2015_239_R_0011, accessed 19 February 2022. Council of the EU (2015b) Council Decision (EU) 2015 2015/1601. Brussels, 24 September, https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32015D1601, accessed 19 February 2022. AMADIO VICERÉ and ANGELUCCI 15 Dahl, R. A. (1971) Polyarchy: Participation and Opposition (New Heaven: Yale University Press). Debomy, D. and Tripier, A. (2017) European public opinion and the EU following the peak of the migration crisis. Policy Paper 201, Jacques Delors Institute, 4 July, https://insti- tutdelors.eu/wp-content/uploads/2018/01/europeanpublicopinionandtheeu-debomy- june2017.pdf, accessed 19 February 2022. Druckman, J.N. and Jacobs, L.R. (2011) Segmented Representation: The Reagan White House and Disproportionate Responsiveness. In P. Enns, and C. Wlezien (eds.) Who Gets Repre- sented? New York: Russell Sage Foundation, pp. 166–188. Ecker, A., Jenny, M., Müller, W. C. and Praprotnik, K. (2022) How and why party position es- timates from manifestos, expert, and party elite surveys diverge: A comparative analysis of the ‘left–right’ and the ‘European integration’ dimensions. Party Politics 28(3): 528– 540. Erikson, R. S. (2013) Policy Responsiveness to Public Opinion. Oxford Bibliographies. DOI: 10.1093/OBO/9780199756223-0103. Fabbrini, S. (2014) The European Union and the Libyan Crisis. International Politics 51(2): 177-195. Gilens, M. and Page, B.I. (2014) Testing Theories of American Politics: Elites, Interest Groups, and Average Citizens. Perspectives on Politics 12 (3): 564–581. Grande, E., Schwarzbözl, T. and Fatke, M. (2019) Politicizing immigration in Western Europe. Journal of European Public Policy 26 (10): 1444-1463. Green-Pedersen, C. and Otjes, S. (2019) A Hot Topic? Immigration on the Agenda in Western Europe. Party Politics 25 (3): 424–34. Hakhverdian, A. (2012) The Causal Flow Between Public Opinion and Policy: Government Re- sponsiveness, Leadership, or Counter Movement? West European Politics 35 (6): 1386– 1406. Hix, S. (2018) When Optimism Fails: Liberal Intergovernmentalism and Citizen Representa- tion. Journal of Common Market Studies 56 (7): 1595–1613. Holsti, O.R. (1996) Public Opinion and American Foreign Policy. Ann Arbor: University of Michigan Press. Hurwitz, J. and Peffley, M. (1990) Public Images of the Soviet Union: The Impact on Foreign Policy Attitudes. Journal of Politics 52: 3-28. Jacobs, L.R. and Page, B.I. (2005) Who Influences U.S. Foreign Policy? American Political Sci- ence Review 99 (1): 107–123. Jolly, S., Bakker, R., Hooghe, L., Marks, G., Polk, J., Rovny, J., Steenbergen, M. and Vachudova, M.A. (2022) Chapel Hill Expert Survey trend file, 1999–2019. Electoral Stud- ies 75 (February): 102420. Kaldor, M. and Selchow, S. (2013) The ‘Bubbling Up’ of Subterranean Politics in Europe. Jour- nal of Civil Society 9 (1): 78-99. Katz, R.S. (1997) Democracy and elections. Oxford: Oxford University Press. Klüver H. and Pickup M. (2019) Are they listening? Public opinion, interest groups and gov- ernment responsiveness. West European Politics 42 (1): 91-112. Lax, J.R. and Phillips, J.H. (2012) The Democratic Deficit in the States. American Journal of Political Science 56(1): 148–166. Leeper, T.J. and Slothuus, R. (2014) Parties and Motivated Reasoning. Political Psychology 35: 129-156. Disconnected? 16 Lefkofridi, Z. and Nezi, R. (2019) Responsibility versus responsiveness…to Whom? A theory of party behavior. Party Politics 26 (3): 334–346. Lenz, G. S. (2012) Follow the leader? How voters respond to politicians' performance and poli- cies. Chicago, IL: University of Chicago Press. Lindberg, L.N., and S. A. Scheingold (1970) Europe’s Would-Be Polity: Patterns of Change in the European Community. Cambridge, MA: Harvard University Press. Mair, P. (1995) Political parties, popular legitimacy and public privilege. West European Poli- tics 18 (3): 40-57. Mair, P. (2009) Representative versus Responsible Government. MPIfG Working Paper 09/8, https://cadmus.eui.eu/handle/1814/12533, accessed 19 February 2022. Mair, P. (2011) Bini Smaghi vs. the parties: Representative government and institutional constraints, EUI Working Papers, RSCAS 2011/22, https://cadmus.eui.eu//han- dle/1814/16354, accessed 19 February 2022. Mair, P. (2013) Ruling the Void: The Hollowing of Western Democracy. London: Verso. McLaren, L. (2003) Anti-immigrant prejudice in Europe: contact, threat perception, and pref- erences for the exclusion of migrants. Social Forces 81 (3): 909–936. McLaren, L. and Johnson, M. (2007) Resources, Group Conflict and Symbols: Explaining Anti-Immigration Hostility in Britain. Political Studies 5 5(3): 709-732. Miller, W.E. and Stokes, D.E. (1963) Constituency Influence in Congress. American Political Science Review 57 (1): 45–56. Miller, W.E., Pierce, R., Thomassen, J., Herrera, R., Holmberg, S., Esaiasson, P. and Wessels, B. (eds) (1999) Policy representation in western democracies. Oxford: Oxford University Press. Moravcsik, A. (1997) Taking Preferences Seriously: A Liberal Theory of International Politics. International Organization 51 (4): 513–53. Moravcsik, A. (2002) In Defense of the ‘Democratic Deficit’: Reassessing the Legitimacy of the European Union. Journal of Common Market Studies 40 (4): 603–34. Moravcsik, A., and F. Schimmelfennig (2019) Liberal Intergovernmentalism. In: A. Wiener, T. A. Börzel and T. Risse (eds.) European Integration Theory. Oxford: Oxford University Press, pp. 64-86. Page, B.I. and Shapiro, R.Y. (1983) Effects of Public Opinion on Policy. American Political Sci- ence Review 77 (1): 175–190. Peffley, M. and Hurwitz, J. (1992) International Events and Foreign Policy Beliefs: Public Re- sponse to Changing Soviet-U.S. Relations. American Journal of Political Science 36: 431- 61. Powell Jr., G.B. (2000) Elections as instruments of democracy: Majoritarian and proportional visions. New Haven, CT: Yale University Press. Rasmussen, A. (2019) Policy representation in Europe: the 2018 Peter Mair lecture. Irish Po- litical Studies 34 (3): 327-349. Sanders, D. and Toka, G. (2012) Is anyone listening? Mass and elite opinion cueing in the EU. Electoral Studies 32 (1): 13-25. Schmitter, P. (2001) Parties are not what they once were. In: L. Diamond and R. Gunther (eds.) Political parties and democracy. Baltimore: Johns Hopkins University Press, pp. 67-89. Simonovits, G., Guess, A.M. and Nagler, J. (2019) Responsiveness without Representation: Evidence from Minimum Wage Laws in US States. American Journal of Political Science 63 (2): 401–410. AMADIO VICERÉ and ANGELUCCI 17 Slothuus, R. and Bisgaard, M. (2021) Party over Pocketbook? How Party Cues Influence Opin- ion When Citizens Have a Stake in Policy. American Political Science Review 115(3): 1090-1096. Steenbergen, M. R., Edwards, E. E. and de Vries, C. E. (2007) Who’s Cueing Whom?: Mass- Elite Linkages and the Future of European Integration. European Union Politics 8(1), 13– 35. Stimson, J.A. (1991) Public Opinion in America: Moods, Cycles, and Swings. Vol. 14. Boulder, CO: Westview Press. Stimson, J.A., MacKuen, M.B. and Erikson R.S. (1995) Dynamic Representation. American Political Science Review 89 (3): 543–565. Thomassen, J. (1994) Empirical research into political representation: failing democracy or failing models? In: M.K. Jennings and T.E. Mann (eds.) Elections at home and abroad, es- says in honor of Warren Miller. Ann Arbor: Michigan University Press, pp. 237-265. Urso, O. (2018) The politicization of immigration in Italy. Who frames the issue, when and how. Italian Political Science Review/Rivista Italiana di Scienza Politica 48 (3): 365-381. Van der Brug, W., D’Amato, G., Ruedin, D. and Berkhout, J. (2015) The Politicisation of Migra- tion. London: Routledge. Van Spanje, J. (2010) Contagious Parties Anti-Immigration Parties and their Impact on Other Parties’ Immigration Stances in Contemporary Western Europe. Party Politics 16 (5): 563–86. Vrânceanu, A. and Lachat, R. (2021) Do parties influence public opinion on immigration? Ev- idence from Europe. Journal of Elections, Public Opinion and Parties 31(1): 1-21. Wlezien, C. (2016) Public Opinion and Policy Representation: On Conceptualization, Meas- urement, and Interpretation. Policy Studies Journal 45 (4): 561-582.