8 - Government Composition and Domestic Conflicts

Summary

This chapter highlights the importance of government composition in explaining the nature of domestic conflict in the refugee crisis. It puts into evidence two important aspects of this composition: fragmentation and ideology. Most of the governments in our study are coalition governments and therefore should not be treated as unitary actors. The type of governments in charge during the crisis ranges from monolithic single-party governments to fractious grand coalitions. Fragmentation is fairly closely linked to the prevalence of intragovernmental conflict, while the link with ideological distance between the parties appeared to matter less. We confirm that center right governments are more likely to engage in debates centered on immigration with the opposition. However, the general relationship between ideology and partisan conflict is weak. The source of the partisan challenge matters more than the ideological make-up of the government: When the challenge comes from the radical right – and to a lesser extent, from the center right – security–sovereignty–identitarian frame types are more likely to be prevalent compared to challenges from the mainstream left, where humanitarian–solidaristic–democratic themes are likely to take center stage.

Introduction

In the previous two chapters, we have offered an analysis on the conflict lines that emerged at both polity levels: between member states and EU institutions, within member states, and between domestic actors and the EU. In this chapter, we continue this line of inquiry by zooming in on two types of conflicts: the conflicts within governments and the conflicts between governments and their domestic partisan opposition. For both types of conflicts, we put the role of government composition at center stage and argue that the fragmentation of governing coalitions as well as the ideological make-up of governments are important determinants of the extent and the type of domestic conflicts that emerge, as well as their substantive content. We consider government composition as an important and yet often overlooked variable in the refugee crisis. Since most of the governments in the countries we study came to power before the crisis reached its peak, their composition can be regarded as largely exogenous to the crisis itself, serving as an overarching constraint on political actors throughout the management of the crisis.

There were exceptions to this rule, however. In the spring of 2017, France experienced a political upheaval as the deeply unpopular government of Francois Hollande was replaced by Emmanuel Macron’s centrist coalition that included ministers from both the traditional left and the right. Later in the same year, one of Austria’s ruling parties, the center-right People’s Party (OVP), ditched its uneasy alliance with the center-left Social Democrats (SPO) and under the new leadership of Sebastian Kurz formed a right-wing government with the FPÖ, Austria’s long-standing radical right-wing challenger party. In the spring of 2018, Italy’s center left Democratic Party, unable to recover from the failed Renzi experiment, was severely punished at the polls and was replaced by the unwieldy populist coalition of the 5 Stars Movement and Matteo Salvini’s right-wing challenger party, the Lega. Finally, Greece also experienced a full-fledged partisan swing from the left to the right: Syriza was defeated decisively at the polls by its conservative rival, New Democracy, in 2019.

Though highly consequential, all these changes came in the later stages of the refugee crisis, which means that most of the policy episodes in our study fell under the departing governments. Moreover, in the four remaining countries, we observe remarkable continuity. After winning the 2015 election with an unexpectedly wide margin, the center right Conservative government of the UK stayed in power, now unconstrained by its previous junior coalition partner, the Liberal Democrats. In Sweden, the center left coalition led by Stefan Löfven came to power just before the start of the crisis and stayed there until the bitter end despite repeated attacks from the right-wing opposition for not taking a harder line against the influx of refugees. Angela Merkel’s grand coalition also survived the crisis despite the highly fractious relationship among the coalition partners, and despite an intervening election in fall 2017, as we shall see in greater detail later on in this chapter. Finally, the crisis did little to dent the stability of Viktor Orbán’s single-party government in Hungary; if anything, it allowed him to tighten his grip over Fidesz and catalyze Hungary’s descent into autocratic rule.

Behind these (partial) continuities within individual countries, however, there is important variation in government composition across policy episodes. In this chapter, we shall assess the explanatory power of this variation in order to account for the type of domestic conflicts that emerged. The first task of this chapter is descriptive: For both intragovernmental and partisan conflicts, we distinguish between various subtypes, relying on the fine-grained information that our PPA dataset provides on the general and specific institutional categories of the actors. Second, we aim to relate various aspects of government composition – namely, government fragmentation and their ideological make-up – to the type of conflict lines. Since our sample is rather limited – forty episodes in total – we limit ourselves to bivariate correlations rather than full-fledged multivariate statistical models, so we lay no claim on any definitive causal link behind the relationships we uncover. Third, we illustrate some of the patterns we have found via episode-specific narratives that illustrate the two main types of domestic conflicts and some of their subtypes. We motivate these empirical exercises, however, with some theoretical considerations derived from the coalition and issue competition literatures in the next section.

Government Composition and Political Competition

As the introductory discussion suggests, the bulk of the refugee crisis was managed by coalition governments. More precisely, twenty-eight out of the forty episodes – in their entirety or during the largest part of their timeline – fell under such government types. The rather obvious observation that coalition governments are not unitary actors has inspired a rich literature in political science, which examines how coalitions are formed (Debus 2008; Laver and Shepsle 1990, 1990; Riker 1984), how they allocate portfolios between each other (Fernandes, Meinfelder, and Moury 2016), and how constituent parties monitor coalition partners to prevent ministerial drift (Indridason and Kristinsson 2013; Martin and Vanberg 2004; Thies 2001). Underneath all these accounts, the common problem that coalition partners need to overcome is the multidimensional and often conflicting objectives they face when they are in government. The classic study on coalition behavior by Müller et al. (1999) distinguishes between three such objectives: policy, office, and votes.

While policy-seeking and vote-seeking behavior by coalition partners potentially pulls them apart as a function of the difference between their policy preferences (ideology) and the preference distribution and the overlap between their electorates, office-seeking motives exert a centripetal force on coalition partners because they have a joint interest in ironing out their differences in order to avoid a government collapse and present a united front to voters as viable coalition partners for the future. Since voters do not assess parties merely for their programmatic and ideological appeals but also for their role and performance as coalition partners (Blais et al. 2006), incentives to signal agreement even against ideological preferences may serve the vote-seeking incentives of coalition partners as well. At the same time, however, coalition partners may also have an incentive to signal disagreement to facilitate voters’ responsibility attribution for policy outcomes (Duch, Przepiorka, and Stevenson 2015) and to counteract voters’ tendency to mesh the ideological profiles of coalition parties by putting them into the same basket (Fortunato and Stevenson 2013). Which specific incentive structure prevails is a highly complex outcome of the party system, the most salient issue area of the day, and the electoral standing of the constituent parties. A complete analysis of all these considerations lies beyond the aim and empirical feasibility of this chapter. We limit ourselves instead to two aspects of government composition as explanatory factors: fragmentation and ideological composition.

The role of government fragmentation is a central insight behind the common pool perspective in budgeting, which argues that with an increasing degree of government fragmentation, the incentives of individual members to internalize the costs and to limit the adverse consequences of excess budgetary demands decrease (Martin and Vanberg 2013; Perotti and Kontopoulos 2002; Roubini and Sachs 1989). We carry this logic forward to intragovernmental conflict beyond budgetary demands and argue that fragmentation within the cabinet is likely to increase incentives by coalition parties to emphasize their differences from coalition partners and reduce incentives to prioritize coalition unity and survival. Such conflict of interest can be especially sharp when coalition partners have equal or comparable access to policymaking levers (Bojar 2019).

The preceding discussion has been ideology-blind in the sense that fragmentation was conceptualized only in numerical terms. Fragmentation, however, has an ideological dimension, too: When coalition members hail from different party families, they are likely to have different policy preferences on immigration and therefore their policy-seeking preferences in the Muller and Storm framework will collide. By contrast, if coalition partners come from the same (or ideologically adjacent) party families, their policy differences are likely to be relatively small, so policy compromise (and lower levels of conflict) is easier to achieve. The second, ideological dimension of government composition thus predicts that with greater ideological distance between coalition partners, intragovernmental conflict is likely to intensify.

The pressure on government parties, as we have seen in the Chapter 6, more often comes from the opposition that tries to pin the government into a corner either by accusing it of doing too little in coming to terms with refugee flows or of excesses and inhumane treatment of refugees. The ultimate source of such partisan conflict is the radical right opposition that has had an immense influence on immigration-related policies over the past decades either directly (Akkerman 2012; Schain 2006; Carvalho 2013) or by putting and keeping the issue on the agenda and compelling government parties to respond by getting tough on immigration both in rhetoric and in substance (Green-Pedersen and Otjes 2019; Bale 2003; Meguid 2005).

Though the distinction between mainstream parties and radical right challenger parties is analytically useful in this regard, we need to take a step further and distinguish between the center-left and the center-right both in government and in the opposition. The distinction is important when one considers the different strategies parties have when faced with issue competition from opposition parties that own an issue that is salient among the electorate (Green-Pedersen and Mortensen 2015). One of these strategies is issue avoidance, as documented in the Swedish context by Odmalm (2011): When parties are faced with challenges from parties that own the immigration issue, it might be electorally worthwhile for them to avoid engaging with the challenge, lest it divert attention from the parties’ core competencies. This consideration is expected to weigh particularly heavily in the calculus of center-left party strategists, which have an ideological inclination to offer a comparatively permissive stance on refugees that may clash with the vote-seeking objectives of the party if forced to compete on the immigration issue. The center-right, by contrast, is comparatively well positioned to compete on immigration (Pardos-Prado 2015), as many of its voters share some of the underlying anti-immigration attitudes that allowed the radical right to capitalize on the refugee crisis (see Chapter 4). Therefore, when center-right governments are in power, partisan conflict is likely to be stronger because governments may actively compete on immigration, either by accommodating the radical right’s demands or by confronting these demands with an emphasis on their own competence to deal with immigration. In sum, our main expectation regarding partisan conflict is that the ideological composition of governments is related to the degree of partisan conflict, with center right governments engaging in more conflictual policy debates with opposition parties than center left governments do.

The foregoing considerations referred to only the intensity of the conflict, not its substantive content. In principle, the conflict both between government actors and between government actors and the opposition can revolve around either overly permissive or overly restrictive immigration policies. Though most of the policies we study imply significant tightening of the countries’ immigration regimes (see Chapter 5), governments can be under simultaneous pressure for breaching human rights and democratic principles and for not going far enough in limiting refugee flows. We expect the ideological composition of the government to be related to whether conflict revolves around humanitarian, solidaristic, and democratic considerations or around securitization, sovereignty-based, and identitarian principles. Specifically, while center left governments are more likely to engage in conflicts on the former grounds, their center right counterparts are more likely to engage in and respond to conflicts revolving around the latter.

Finally, in terms of partisan conflicts, it is not just the ideological composition of the government that matters but also the origin of the conflict. When conflicts emerge between the government and its right-wing opposition (either center right, or radical right), the security–sovereignty–identity mix is likely to predominate when compared to conflicts that emerge between governments and their left-wing opposition.

Data and Measurement

Many of the variables we use to test our theoretical expectations are based on the PPA dataset that we use throughout the book. In order to measure the intensity of intragovernmental and partisan conflicts, we revert to the conflict scores we derived in Chapter 6. In this chapter, we focus on only the intragovernmental and the partisan conflicts. We shall further investigate which particular actor pairs contribute most to these two conflicts. Within intragovernmental conflicts, the debate can unfold according to four different scenarios: within governing parties, between coalition partners, between government parties and the government, and within the government itself (for instance, between the prime minister and particular ministries). As for partisan conflicts, one of the conflicting parties is always the government (or government parties), but the adversaries can be the radical left, the radical right, the mainstream left, or the mainstream right. Figures 8.1 and 8.2 illustrate the distribution of these conflict sources in the policy episodes that we classified as intragovernmental conflicts and partisan conflicts, respectively.

Figure 8.1 The sources of intragovernment conflicts in the refugee crisis

Figure 8.2 The sources of partisan conflicts in the refugee crisis

Figure 8.1 shows the relative distribution of the four sources of intragovernment conflicts. Overall, the most common source is conflicts between government parties and the government, which arguably reflects the fact that parliamentary actors sought to achieve some sort of oversight over the crisis management of what has been predominantly an executive affair. In fact, more than half of such party–government interactions were initiated by senior government parties and targeted at the government. Comparatively speaking, conflicts within the government were rarer, on average. However, such conflicts were the dominant sources of intragovernmental conflicts in the Sicurezza Bis episode in Italy. In this episode, such within-government conflict was a triangular debate between the prime minister (Giuseppe Conte), the interior minister (Matteo Salvini), and the ceremonial head of state of the Italian Republic (Sergio Mattarella). Such a premier–interior minister stand-off was replicated in the CDU-CSU Conflict in Germany, where Chancellor Angela Merkel and Interior Minister Seehofer got caught in an acrimonious debate over the latter’s emboldened push toward a tighter asylum regime and an accelerated deportation process after becoming interior minister in the autumn of 2018.

Some degree of interparty debate was present in five out of the six intragovernmental conflicts, but in none of them was it particularly intense, with the partial exception of the integration law debate in Germany, where the three coalition partners – CDU, CSU, and SPD – exchanged verbal blows, with the SPD taking the leading role by criticizing the CSU on its hardline stance. Conflicts within the parties themselves were least common. They appeared only in the German episodes, whereas the Austrian and the Italian government parties managed to maintain party discipline and concentrated their efforts on criticizing coalition partners or the government.

Turning to partisan conflicts, Figure 8.2 displays their sources. As the reader may recall, partisan conflicts are significantly more common than intragovernmental conflicts, and there is a larger variation in the partisan patterns. What is immediately apparent is that governments engage in conflict much more often with their mainstream opposition rivals (especially with the center right) than with their radical challengers. The role of radical left challengers is especially limited. The mainstream right is an important source of conflict in six out of the thirteen episodes – in three of the four French episodes, in the Summer of 2015 episode in Greece, in one of the Hungarian episodes (“Stop Soros”¹) and the two Swedish partisan conflict episodes. The fact that the mainstream right has been a more vocal opponent of governments than the mainstream left provides early tentative support for the expectation that the center right has more to gain from politicizing immigration than the center left does.

On the radical end of the partisan spectrum, the dominance of the radical right is unsurprising. It has been the most vocal opponent of governments in one Austrian (Right to Intervene) and two French (Border Controls and Calais) episodes, consistent with their long-established presence in the political scene of the two countries. Comparatively speaking, Golden Dawn, Jobbik, the Sweden Democrats, and UKIP have accounted for a much more limited share of partisan conflict with the Greek, Hungarian, Swedish, and British governments. Overall, the share of a conflict that is attributable to the radical right tends to be higher in contexts where it is electorally stronger, such as France and Austria. The correlation coefficient between the average electoral strength of the radical right challengers throughout the policy episodes and the share of the partisan conflict with the radical right is 0.38. By contrast, the participation of the radical left in government–opposition conflicts is restricted to five of the thirteen episodes, and in none of them did it become a particularly prominent feature of the debates. The only partial exception is the Rights of Foreigners bill in France, but even here, merely two actions were targeted at the government by radical left politicians from the New Anti-Capitalist Party and the Radical Left Party.

After this brief overview of the sources of intragovernmental and partisan conflicts, we now return to the variables we highlighted as potentially important explanatory factors for the strength and substantive content of the conflicts.² For government fragmentation, we use the Herfindahl Index of governments from the Database of Political Institutions (Cruz et al. 2021), which measures the sum of the squared seat shares of all parties in the government. In case of single-party governments, this indicator takes a value of 1, whereas for large coalitions constituted by many parties of roughly equal strength, it is close to 0. In our sample, none of the governments were particularly fragmented, so the effective distribution of the variable in our sample is situated between 0.5 and 1.

For the ideological variable, we rely on the GALTAN (Green–Alternative–Libertarian, Traditional–Authoritarian–Nationalist) score of parties assigned by experts participating in the Chapel Hill Survey (Jolly et al. 2022). We use the respective scores from the survey wave closest to the corresponding policy episodes. The GALTAN score locates parties on a 0–10 scale, with higher values assigned to parties taking a position closer to the Traditional–Authoritarian–Nationalist pole and lower values for positions closer to the Green–Alternative–Libertarian pole of the attitudinal divide (Hooghe et al. 2002). We measure the ideology of governments by the average of the governing parties, weighted by their seat shares in parliament. For ideological fragmentation, we take the average absolute distance between the GALTAN scores of the governing parties.

As Figure 8.3a reveals, governments in the refugee crisis spanned the entire ideological spectrum, with a slightly rightward skew. The most ideologically right-wing government (the third Orbán government in Hungary) is closer to the TAN pole than the most left-wing ones (the Renzi/Gentiloni governments in Italy) are to the GAL pole. Moreover, fifteen of the forty episodes occurred under left-of-center, and twenty-five occurred under right-of-center governments. The typical form of such left-of-center governments was a coalition between left-wing parties. An example of this constellation is the Swedish case, where the Social Democratic Party was in a coalition with the Green Party throughout all five Swedish policy episodes. Among right-of-center governments, we observe two main types. Twelve of the twenty-five right-of-center governments were single-party governments, such as the Fidesz-led governments in Hungary and the Mitsotakis-led government in Greece during the late Greek episodes in the years of 2019 and 2020. Another twelve were grand coalitions, which, due to the ideological position of the constituent parties as well as their relative strength, score above 5 on the weighted ideological position variable. Examples of such right-of-center grand coalitions are the German and the Austrian grand coalitions as well as the Lega–M5S government in Italy. In fact, the only left-of-center grand coalition in our sample is the French government led by President Macron’s centrist REM party during the Asylum Law episode. As for the ideological distance (displayed in the lower panel of Figure 8.3), single-party governments score 0, by construction. Most governments’ ideological distance varies in a moderate range between 0.2 and 0.4, and only a few governments display large differences between the coalition members on the GALTAN scale. This group includes two Greek episodes under the Syriza–Anel coalition and two Italian episodes under the M5S–Lega coalition.

Figure 8.3 Ideological position (a) and distance (b) of governing coalitions in the refugee crisis

The final measurement issue concerns the substantive part of the conflict. To this end, we rely on frame scores in our PPA coding, which distinguishes between ten frames actors use to justify their position/action (see Chapter 9). We distinguish between security–sovereignty–identitarian frames on one end and humanitarian–solidarity–democratic frames on the other. Our measure for the substantive part of the debate is then the share of these two types of frames among all the frames used. We limit this calculation to those actions that constitute the respective conflict lines for intragovernmental and partisan conflicts.

Government Composition and Political Conflict in the Refugee Crisis

We begin the empirical investigation with the relationship between government fragmentation and intragovernmental conflicts³. As Figure 8.4 shows, the relationship is in the expected direction. All six episodes where such intragovernmental conflicts predominate are characterized by high levels of government fragmentation (relatively low scores on the Herfindahl index). On the other end, episodes falling under single-party governments all have a lower than average intragovernmental conflict score. The correlation between the two variables is rather high (–0.61), and even if we exclude all single-party governments from the sample and concentrate on coalition governments only, it is not much lower (–0.58).

Figure 8.4 Government fragmentation and intragovernmental conflicts

It must be emphasized, however, that the high levels of intragovernmental conflict associated with government fragmentation are largely driven by the German and the Austrian grand coalitions, as is readily visible in Figure 8.4. Though we operationalized government fragmentation simply by the relative strength of the constituent parties, certain other idiosyncratic features of these grand coalitions beyond party fragmentation provide equally important parts of the story. In the German case, one of these features is the role of the CSU, the Bavarian sister party of the senior government party, the CDU. A significant part of the intragovernmental conflict played out between this regional party and Chancellor Merkel and her party as well as the junior coalition member, the SPD. The leader of the CSU, Horst Seehofer, who also became interior minister in March 2018, played an especially pronounced role in this conflict due to his hardliner stance against Chancellor Merkel’s “Willkommenskultur” (see Chapter 6). In the Austrian case, an important venue for this intragovernmental conflict was the interaction between regional authorities and the central government. Although in Chapter 6 we treated such interactions as a distinct state–government conflict, it is important to recognize that several regional politicians played a prominent role in one of the government parties, such as Hans Niessl, SPÖ governor of Burgenland; Josef Pühringer, ÖVP governor of Upper Austria; and Michael Häupl, SPÖ mayor of Vienna. From their position as regional politicians, therefore, they also contributed to intragovernmental conflicts, particularly in the Asylum Law episode, where they launched no fewer than fourteen critical actions against the federal government. Intragovernmental conflicts in grand coalitions can thus be conceptualized as a result of government fragmentation in a broader sense that includes fragmentation across different levels of policymaking, particularly in federal countries.

The correlation between ideological fragmentation (average ideological distance between the coalition partners) and intragovernmental conflict score is considerably weaker, albeit still in the expected direction: 0.27. This relationship is, however, largely driven by single-party governments, where the ideological distance is zero by definition, as we saw on Figure 8.4, and which tend to be characterized by low intragovernmental conflicts. When focusing on coalitions only, the correlation coefficient is a mere 0.06, which provides very limited evidence for our expectation that ideological distance is a determinant for conflict. A more plausible interpretation of the data is that government fragmentation is a likely determinant of intragovernmental conflict even if coalition partners hail from similar or ideologically proximate party families. This is likely to be the result of the fact that conflict over immigration within the government is not necessarily a result of different ideological principles but rather follows from debates over electoral strategies, policy details, or blame avoidance strategies by the coalition partners.

A case in point is the Integration Law episode in Austria. This episode has the highest intragovernmental conflict score in the whole sample, and it occurred under a grand coalition government with a relatively large ideological distance (3.2) between the two constituent parties, the center left SPÖ and the center right ÖVP. In the conflict, however, few of the actions emanated from distinct ideological principles. The most contentious elements of the debate revolved around a ban on veiling in public places, language requirements, and a requirement for refugees to accept nonprofit jobs. Despite the sensitivity and ideological divisiveness of these issues, few of the conflictual actions revolved around basic ideological principles; rather, they focused on procedural matters and took the form of the parties mutually accusing each other of not sticking to their part of the coalition bargain. What this anecdotal evidence suggests is that even in cases with a relatively large ideological distance between coalition partners, the debate between them can be rather nonideological, so government fragmentation alone is a sufficient condition for intragovernmental conflicts.

Turning to the second conflict dimension that we examine in this chapter, we probe the relationship between the ideological make-up of governments and the extent and type of partisan (government–opposition) conflict (see Figure 8.5). As it turns out, the relationship between the average ideological position of the government on the GALTAN scale and the partisan conflict score of the episodes is positive but weak (0.2). On the one hand, with one exception, governments scoring high on the GALTAN scale (>8) are characterized by comparatively high levels of partisan conflict. On the other hand, all but two of the governments scoring low on this scale (<4) produced below-average partisan conflict scores. The two outliers among the governments with low scores – the Hollande government during the Rights of Foreigners Bill in France and the Swedish center left coalition led by Löfven in Sweden during the Residence Permits episode – were both characterized by a partisan context where the government was simultaneously attacked from both the left and the right. Though in both cases, the mainstream center right opposition led the offensive, the government was also criticized by the radical right and the radical left opposition. What explains the outlier status of these cases, therefore, is the multiple angles of partisan attack against the government rather than the government proactively seeking out conflict.

Figure 8.5 Government ideology and partisan conflicts

We expected that the center-right, when in government, may have a lot to gain from politicizing immigration, in contrast to the center-left. Though it is highly questionable to what extent Viktor Orbán’s Fidesz party can be considered a center-right party, the location of the two Hungarian partisan conflicts – the Quota referendum and the “Stop Soros” episode – in the upper right quadrant of Figure 8.5 suggests that these two episodes provide a useful testing ground for the validity of this mechanism. If our expectations are valid, the government (and government parties) should be in the driver’s seat as instigators of the conflict. This is definitely the case for the “Stop Soros” episode, where the government (and its parliamentary wing) was responsible for almost half (47 percent) of conflictual actions. If the government’s initiating role in the Quota Referendum episode was less pronounced, accounting for 36 percent of the partisan conflict, it is still situated well above the average government share for all partisan conflicts (roughly a quarter). Overall, however, as the proponents and the executors of the policy packages, governments are obviously much more likely to be targets rather than initiators of the conflict when forced to defend their policy proposals (see Chapter 6). Against this backdrop, the two Hungarian cases, particularly the “Stop Soros” episode, provide some suggestive evidence for the prospects of the center-right to electorally benefit from putting the immigration issue on the political agenda.

We also anticipated that government ideology determines the conflicts’ substantive content, as it is conveyed to the audiences via particular framing strategies that actors employ in the debates. In particular, we expected that center-left governments are more likely to engage in debates on humanitarian–solidarity–democratic grounds, whereas the center-right would prioritize debates revolving around themes related to security, sovereignty, and cultural (identitarian) concerns. However, the correlations between the share of such frame types and government ideology provide only weak support for this argument. For intra governmental conflicts, the relationship between government ideology and the share of humanitarian–solidaristic–democratic frames is indeed negative (correlation coefficient: –0.26), while the share of security–sovereignty–identitarian frames is positive (correlation coefficient: 0.17). For partisan conflicts, however, the resulting patterns go against expectations, with correlation coefficients for the types of frames amounting to, respectively, 0.17 and 0.06. This unexpected result can be explained by the fact that in the case of partisan conflicts, governments have only partial control over the substance of the debates. For instance, left-wing opposition parties may trigger conflicts on humanitarian grounds even if the center right government tries to ignore their actions. Conversely, center left governments may decide to ignore conflicts around security concerns, but that does not make the conflicts go away, as (center) right opposition parties can keep such security threats on the agenda against the wishes of the government.

To test this proposition, we now turn to the final empirical exercise, which relates the substantive content of the conflict to the sources of partisan conflict. To reiterate our expectations, we expected the share of humanitarian–solidaristic–democratic frames to be higher when the main opposition challenge comes from the left, and we expected the share of security–sovereignty–identitarian frames to be higher under challenges from the right. Starting with the share of the first type of frames, the patterns are closely in line with our expectations: The correlation coefficient between the share of such frames and the share of partisan conflict emanating from the center left and the center right are 0.63 and –0.40, respectively. For the conflict between radical challengers, the coefficient is –0.35 with radical right challengers and exactly zero (no correlation) with radical left challengers. The lack of correlation with the radical left is most probably related to the very low number of such actions in our sample. For frame shares of the security–sovereignty–identitarian type, the patterns largely mirror the previous findings, with one important difference. For such types of frames, it is the radical right that appears to be successful in putting such concerns on the agenda (the correlation coefficient between the share of these frames and the share of the conflict originating from the radical right is 0.52). The corresponding correlation coefficient for the mainstream left amounts to –0.38, for the mainstream right it is –0.04, and for the radical left it is –0.02 (again, no correlation). Therefore, while the center right has been a more vocal opponent of governments during the refugee crisis, the radical right opposition has been more successful in sticking to a consistent securitarian and sovereigntist narrative in the debate.

To sum up the findings on the relationship between the substantive content of the conflict and the source of partisan challenges, the most noteworthy patterns are the high share of humanitarian–solidaristic–democratic frames when the opposition comes from the center-left and the high share of security–sovereignty–identitarian frames when the challenge comes from the radical right. The role of the center-right is somewhat ambiguous: Though its presence, when in opposition, is associated with less conflict on humanitarian–solidaristic–democratic grounds, this does not translate into a higher share of security–sovereignty–identitarian concerns. More generally, we have seen that while the ideological composition of the government is consistent with the substantive content of the conflict in the intragovernmental domain, it is the ideological source of partisan opposition that has a stronger predictive power related to the content of the debate in the government–opposition (partisan) domain. Figure 8.6 shows the scatterplots with the highest correlation coefficients in such partisan conflicts.

Figure 8.6 Relationship between the content of the conflict and their partisan source

The Impact of Government Composition in Action: Two Case Studies

Sicurezza Bis in Italy (September 2018 to August 2019)

Based on the composition of the government that presided over the Sicurezza Bis episode in Italy toward the end of the refugee crisis, the episode was always going to be a perfect candidate for intragovernmental conflict. The coalition was composed of two main parties with vastly different ideological profiles on the GALTAN dimension: the Lega, an archetypical populist radical right party and M5S, a relatively new actor on the Italian party scene with a rather motley ideological profile but as far as the cultural dimension is concerned, arguably playing the role of the functional equivalent of a new left party in the Italian political system (Kriesi 2020). Government fragmentation was thus rather high both in numerical terms (0.54 on the Herfindahl index) and in terms of the ideological distance between the parties (5.47). Moreover, the M5S–Lega government was a case of nonaligned setting between the interior portfolio and the prime minister. Arguably, the interior minister and the leader of the Lega party, Matteo Salvini, even eclipsed the role of the nonpartisan premier Giuseppe Conte in this episode.

Unsurprisingly against this backdrop, the episode turned out to be one of the six intragovernmental conflicts and the only one that emerged outside Germany and Austria, the two countries with long-standing traditions of grand coalitions and the inevitable conflicts these entail. Most of the conflict played out within the government itself, as we have briefly mentioned before. Thirteen of the fifteen intragovernmental exchanges took place between government actors, while the remaining two occurred between government parties and the government. The relative peace between the coalition partners, however, is largely due to the fact that the Lega was largely a one-man show led by Salvini, who now acted in his new role as interior minister, rather than as the head of his party.

Salvini and Conte contributed to the conflict in roughly equal measure both as initiators and as targets. Predictably, they targeted each other most of the time. Though Salvini refrained from outright criticism of the premier and relied on softer forms of pressure via a radio interview, a letter directed at him, and statements made in a government meeting, his actions were largely aimed at speeding up the process of approving the law that sought to tighten the asylum system by accelerating deportations and facilitating the detainment of asylum seekers. In exchange, Conte expressed doubts on the constitutional legality of the decree and invoked the president of the republic, Sergio Mattarella, who shared these concerns.

In fact, Mattarella pushed his constitutional prerogatives to the limits by expressing concerns about the decree on various occasions. In early October, he invited Salvini for a meeting in the Quirinale – the Italian presidential palace – to express reservations about the law. Later, in a letter addressed to the government, he emphatically demanded that the constitutional rights of foreigners be respected. Much later, at the end of the episode, he made a last-minute attempt to curb the excesses of the law in yet another letter addressed to the leaders of both chambers, where he labeled the sanctions of those violating territorial waters “unreasonable,” a rather harsh expression from the president in an otherwise civilized debate.

Amidst the Salvini–Conte–Mattarella triangle, the role of Luigi di Maio and the senior coalition party, M5S, was somewhat ambiguous. Though he sought to assuage the concerns of Salvini by promising that he would impose order in the ranks of his party and get the votes to support the decree, at the same time, he did not shy away from distancing himself from the interior minister. In one statement, he accused the latter of trying to push the decree through without proper consultation with his party: “Salvini is trying to provoke us to cover up his failures, we will not fall into the game of responding to a decree that no one has ever discussed in advance.” In an inner-circle discussion, he went even further by accusing Salvini (and the Lega) of threatening the survival of the government and at the same time thought to assuage his party, saying that he would not give in to all of Salvini’s demands. Ultimately, however, this balancing act of di Maio turned out to be a failure because the substance of the decree ended up largely representing the Lega’s (and the populist radical right’s in general) vision of clamping down on asylum seekers in the context of the crisis.

In terms of the substance of the debate and the frames that the actors used, the main patterns also largely conform to our theoretical expectations. We argued that center right governments are more likely to engage in intragovernmental conflicts on security–sovereignty–identitarian grounds, whereas center left governments would prioritize humanitarian–solidaristic–democratic concerns as far as the intragovernmental conflict line is concerned. The ideological placement of the M5S–Lega coalition government is far from trivial because of the ideological ambiguity of M5S. The Chapel Hill expert survey scores place the Lega firmly on the right of the GALTAN spectrum, whereas M5S is coded as center left, giving rise to a weighted average ideological score of 5.67 (i.e., slightly right of center) for the government. Considering that both the nonpartisan premier Giuseppe Conte and the head of state Sergio Mattarella, who played a prominent role in the episode, had entered politics from a legal background, the overall weight of the government is expected to tilt further to the center. Accordingly, the frame mix in the debate was rather balanced. In the overall debate, roughly a quarter of the frames are of the security–sovereignty–identitarian mix, and slightly less than half are humanitarian–democratic (no solidaristic frame was used in this episode).

When zooming in on the part of the debate that unfolded along the intragovernmental conflict line, the balance is roughly the same: Two actions were accompanied by a security–sovereignty frame and three by humanitarian and democratic ones. Starting with the security–sovereignty types, both of these actions were undertaken, unsurprisingly, by Salvini. In May 2019, he defended the proposed measures to his followers on the grounds that they would protect Italy against “smugglers, criminals, and convicts,” rhetorically musing about how the coalition partner M5S could possibly be against the proposal. In the same month, in a letter addressed to Premier Conte, he sought to dismiss concerns voiced by six UN rapporteurs, calling these interventions “undue invasions” in a domestic political matter. On the other end of the frame mix, Sergio Mattarella played the leading role yet again. On various occasions, he invoked the constitutional rights of foreigners, and in the letter sent to the heads of the legislative chambers toward the end of the episode, he stressed that “there is always a responsibility to rescue at sea.”

Quota Referendum in Hungary (November 2015 to December 2016)

If the Sicurezza Bis episode in Italy created fertile grounds for intragovernmental conflicts to emerge, the one-year-long Quota Referendum episode in Hungary was an equally likely candidate for partisan conflict. While the party discipline that Viktor Orbán, the Hungarian premier, imposed on his single-party government all but precluded any sort of dissent from the government’s ranks, the ultraright policy platform of Fidesz (GALTAN score: 8.6) foreshadowed that not only would the government be exposed to frequent attacks from the opposition but that it would also readily engage with such attacks or provoke the conflict itself. In fact, as we previously showed, in the two Hungarian partisan conflicts – the Quota Referendum and “Stop Soros” – the government and its parliamentary wing initiated a comparatively large share of the partisan conflict. Moreover, given the rather heterogenous partisan opposition standing against Fidesz, the government’s plans to block the EU’s relocation scheme via a referendum was likely to be criticized from multiple directions and substantive angles. Accordingly, the episode came to be dominated by the partisan conflict line (with a partisan conflict score of 0.40). No less than 123 actions involved the government and the opposition, an outstanding number among our episodes both in relative (as a share of total actions) and in absolute terms.

Most of the government’s attacks were targeted at the long-standing leader of the mainstream left, the postcommunist MSZP (Magyar Szocialista Párt) party. However, by the time the referendum initiative was launched, the radical right challenger party, Jobbik – which, incidentally, initiated the constitutional change to block the relocation of asylum seekers to Hungary in the first place – had overtaken MSZP as the leading opposition force and was steadily climbing in the polls. As Jobbik’s challenge was widely perceived as more threatening to the government than the left-wing opposition, Fidesz could not ignore it and often targeted Jobbik in the debate. From the opposition’s side, the most active initiator of the conflict was yet again MSZP, followed by ex-premier Ferenc Gyurcsany’s Democratic Coalition. Jobbik was comparatively silent as an initiator, not least because the referendum initiative was close to its original plans and its general policy agenda. Nevertheless, Jobbik also targeted the government on eight different occasions. Finally, LMP – a nominally green left outfit but in political terms a centrist party playing a “bridging” role between the two blocs – also participated in the debate, though it was largely spared from the kind of government offensive that other opposition parties had to face.

Apart from its intensity, one of the most unique features of the partisan debate is the multiple arenas in which it unfolded. The media accounted for only around a third of the action, a limited share when compared to other episodes. This is largely due to the fact that in the summer as the date of the referendum (October 2016) neared, the debate gradually shifted to public campaigning, including various poster campaigns and other official campaign events. A number of opposition protest events also took place – including conventional demonstrations as well as a “human chain” around the parliament organized by the Democratic Coalition, most of them immediately before the referendum vote. Finally, the referendum also loomed large in the parliamentary arena, both in the preparatory phase in the spring and in the referendum phase in the autumn, when Fidesz first tried to mobilize the vote to reach the quorum and then to impose a constitutional amendment despite the unsuccessful referendum outcome for its position. Meanwhile, the opposition’s main strategy in the parliamentary debate was to take an ambiguous stance on the Relocation Scheme as such, while arguing that the referendum was a futile tool to fight it. However, there were discernible differences in the strategies between the mainstream opposition and Jobbik. The former sought to highlight the government’s incompetence and hidden agendas while refusing to take a firm stance on the fate of refugees, whereas Jobbik was careful to emphasize its substantive policy agreement with the government even as it criticized the latter on procedural grounds.

Similar to the Sicurezza Bis episode in Italy, the prevalent frames in the policy debate were rather mixed. Conspicuously, solidarity frames, yet again, were entirely absent from the debate, which is somewhat paradoxical given that the debate was ultimately about interstate solidarity. Instead, while the government successfully promoted its own narrative on security and identitarian grounds – with sovereignty frames taking a secondary role, the frame mix by the opposition mostly centered around democratic/legal norms – humanitarian considerations were invoked only once, in a mocking response to the government’s poster campaign. Instead, the most common frame type employed by the opposition was one of efficiency/pragmatism. This conforms to its overall strategy that we highlighted before: Instead of attacking the government on principled grounds, the opposition mostly aimed to highlight the futility of the referendum push. The relatively low share of security–sovereignty–identitarian frame types thus partly goes against our expectation that such frames should prevail if the mainstream left is the main source of conflict. Even if somewhat unexpected, this outcome can be accounted for by the government’s successful dominant initiating role in the conflict and the support it obtained from Jobbik, ever so careful to emphasize its toughness on immigration.

Conclusion

This chapter has highlighted the importance of government composition in explaining the nature of domestic conflict in the refugee crisis. We have put into evidence two important aspects of this composition: fragmentation and ideology. Our focus on government fragmentation was informed by the notion that most of the governments in our study are coalition governments and therefore should not be treated as unitary actors. The type of governments in charge during the crisis ranges from monolithic single-party governments – such as the Fidesz government in Hungary and the Mitsotakis government in Greece – to fractious grand coalitions. Some of these coalitions are further fragmented on ideological grounds, as we have witnessed in the case of the M5S–Lega coalition in Italy.

Our empirical exercises relying only on bivariate correlations due to the limited sample size in our study revealed some interesting patterns regarding the relationship between government fragmentation and the intensity of the intragovernmental conflict line. Numerical fragmentation showed a fairly close link to the prevalence of this conflict, while the link with ideological distance between the parties appeared to matter less. Ideology turned out to be of mixed relevance for the intensity of the partisan conflict. We confirmed that center-right governments are more likely to engage in debates centered on immigration with the opposition. Moreover, the role of government ideology also matters for the content of the debate along both the intragovernmental and the partisan conflict lines. However, the general relationship between ideology and partisan conflict is weak.

In substantive terms, we expected (and empirically confirmed) that center-right governments are more likely to engage with immigration-related debates among themselves compared to center left governments, whose electoral incentives push them to hide their differences and emphasize other issues instead. However, when the conflict unfolds between the government and the opposition, we have seen that the source of the partisan challenge matters more than the ideological make-up of the government: When the challenge comes from the radical right – and to a lesser extent, from the center right – security–sovereignty–identitarian frame types are more likely to be prevalent compared to challenges from the mainstream left, where humanitarian–solidaristic–democratic themes are likely to take center stage. In practice, however, as we have seen both in the case of the Sicurezza Bis episode in Italy and in the case of the Hungarian Quota Referendum, the frames that dominate the debates tend to be highly mixed and variegated, and they are likely to depend on a host of other factors beyond ideology and the general scope of this chapter.