The conveyance and effect of the asylum abuse policy narrative in direct- democratic campaigns

David Kaufmanna & Laurent Bernhardb

a KPM Center for Public Management, University of , [email protected]

b Swiss Centre of Expertise in the Social Sciences (FORS), University of Lausanne,

[email protected]

Abstract

In public and policy debates, asylum seekers are often portrayed as ‘bogus’ refugees or economic migrants who are accused of trying to abuse the country’s generosity and protection.

We make use of the Policy Narrative Framework to systematize this accusation. We study the conveyance of the abuse policy narrative by political elite actors and the effect of this narrative on citizens’ opinion formation in the three most recent direct-democratic campaigns on Swiss asylum policies (2006, 2013, and 2016). We conducted and statistically analyzed over 108 structured interviews with all relevant campaign managers to analyze the conveyance of the abuse policy narrative and we analyzed post-vote surveys to examine the effect of this narrative for citizens’ opinion formation. We find that the more to the right a political organization is, the more this organization relies on the abuse policy narrative but only in referendums about tightening of asylum policies. We find similar results for citizens’ opinion formation as the asylum abuse policy narrative is crucial in tightening reforms for citizens from the center, the moderate right and was well the radical right. This conceptualized asylum abuse policy narrative can be used to study public debates about asylum policies beyond referendum campaigns and the findings offers clues about the use and the effect of the asylum abuse policy narrative in wider political debates.

1

Introduction

The collapse of the Soviet Union, the communitarization of European asylum policy and high numbers of asylum seekers due to crises in close proximity to Europe motivated the formation of elaborated asylum systems in Western Europe (Hatton 2012; Zaun 2017). Periods of high numbers of asylum seekers and the politization of asylum policies by political actors from the radical right have contributed to the salience of asylum policies salient in Western European political systems (Bale 2008; Hatton 2009; Hatton 2012). Asylum policies in Western Europe are characterized by an agitated public, mobilized interest groups and partisan conflict

(Freeman 2006, 238).

In these public debates over asylum policies, asylum seeker are often portrayed as economic migrants and bogus refugees who are accused of trying to abuse the country’s generosity and protection. This popular accusation of abusing national asylum systems can be conceived as a policy narrative. In simple terms, the abuse policy narrative constructs a policy problem (i.e., the abuse of benevolent European asylum systems) and puts forward a policy solution to this constructed problem (i.e., tighter asylum policies). This kind of abuse policy narrative has been prevalent in the policy and public discourses about asylum policies in many asylum destination countries such as in the United Kingdom (Bloch 2000; Sales 2002; Schuster 2004; Schuster and

Solomos 2004; Goodman and Speer 2007; Jennings 2010; Darling 2014), in

(Skenderovic 2009; Hänggli and Kriesi 2010), in the discourses about boat arrivals in Australia

(Zagor 2015) as well as in the discourses about migration more generally in the United States

(McBeth and Lybecker 2018). The distinction between refugees ‘worthy’ of protection and

‘bogus’ (economic) migrants as a central element of the narrative is enshrined in state policy practices as well as popularized in public discourse (De Genova 2002; Darling 2014; Skleparis and Crawley 2018).

In this article, we consider the popular accusation that the majority of asylum seekers abuse national asylum systems as a policy narrative. Policy narratives can be studied by using a 2

heuristic proposed in the Narrative Policy Framework (NPF) that structures policy narratives in four core elements: setting, characters, plots, and moral (Jones and McBeth 2010; Shanahan,

Jones, and McBeth 2011; Shanahan et al. 2017). We contextualize the core elements of the NPF for our study of asylum policies in the Swiss direct-democratic campaigns.

We study the conveyance of the asylum abuse policy narrative of political actors in direct- democratic campaigns on Swiss asylum policies as well as the influence of this narrative on citizens’ opinion formation. In the terminology of Shanahan et al. (2017), we combine a mesolevel with a microlevel analysis. Thereby, we want to contribute to current developments of the NPF as there is a growing interest in understanding the connectedness between the different levels of analysis (Crow 2012; Shanahan et al. 2017, 196). To link a mesolevel analysis of how elite actors convey policy narratives with a microlevel analysis of citizens’ opinion formation is described as an intriguing underdeveloped aspect of the NPF research (Shanahan et al. 2017, 197).

Intensive public debates on asylum policies have repeatedly taken place in the context of Swiss direct-democratic votes (Bernhard 2012: 41-45). In the thirty-year period between 1987 and

2016, eight asylum-related propositions were placed on the ballot at the federal level. Political actors tactically use narrative elements to mobilize citizens to vote in Swiss direct-democratic campaigns (Schlaufer 2018). Thus, these direct democratic campaigns offer an opportunity to study who uses the abuse policy narrative, how it is used in the context of asylum policies, and how it influences citizens’ opinion formation. We thus address the central questions of the NPF, by asking “do narratives play an important role in the policy process?” (McBeth et al. 2014,

225).

The empirical analysis is based on 108 structured interviews with campaign managers in advance of the three most recent referendums regarding federal asylum law in 2006, 2013, and

2016. In addition, we make use of VOX post-vote surveys to assess the effect of the abuse policy narrative on citizens’ opinion formation. We find that the role played by the abuse policy 3

narrative strongly varies across campaigns. More specifically, we find that the political elite actors convey the abuse policy narrative most often in the tightening reform of 2006, followed by the balanced reform of 2013 and the streamlining reform of 2016. The more to the right a political elite actor, the more likely this actor makes use of the abuse policy narrative in referendum campaigns on Swiss asylum policy. Similar results appear in the analysis of the effect of the abuse policy narrative on citizens’ opinion formation: The abuse policy narrative was more important for citizens’ opinion formation in tightening campaigns and it is important for citizens from the center, the moderate right, and the radical right.

Abuse as a policy narrative

Narratives are crucial for politicians, strategist, and media reporters given the importance of how a story is rendered for policy success (Shanahan et al. 2017). The NPF became the decisive framework for studying narratives in the policy process because it allows to study the construction of policy problems and their causes as well as their proposed solutions by means of a plot (Schlaufer 2018). Narratives are perceived as crucial parts of constructing policy realities and, thus, the NPF acknowledges the importance of social construction in the policy process (Stone 1989; Shanahan et al. 2017; Schlaufer 2018).

As an important tool, the NPF offers a general structure of policy narratives that is generalizable over time and space (Jones and McBeth 2010; Shanahan, Jones, and McBeth 2011; Shanahan et al. 2017). This structure contains of four core elements. Frist, the setting accounts for the embeddedness of policy narratives in the specific institutional, geographical, legal, or economic context. Thus, the policy narrative depends on the specific policy context and what is perceived as the policy problem, which in turn relies on the social construction of policies. Second, policy narratives contains at least one character. “As with any good story, there may be victims who are harmed, villains who do the harm, and heroes who provide or promise to provide relieve from the harm and presume to solve the problem” (Shanahan et al. 2017, 176). Third, the plot 4

establishes the relationships between the characters and situates the characters within the policy setting. It serves as the arc of the action within the narrative. And fourth, the policy solution for the policy problems is the moral of the story. This structures allows to analyze and compare policy narratives in different policy contexts (Shanahan et al. 2017, 175-177).

In the following, we use the structure of the NPF to theorize the abuse policy narrative in Swiss direct-democratic asylum campaigns. Given the context of direct-democratic campaigns and the contextual characteristics of the asylum policy subsystem, we discuss the setting element of the NPF structure in detail. The abuse message in itself, with regard to the characters, the plot and the moral of the story, is, however, rather simple.

Overall, the abuse policy narrative employs an intentional causal mechanism narrative1 (Stone

1989; Shanana et al. 2017) since it strategically arranges narrative elements to assign blame to a large population of asylum seekers who are blamed to intentional abuse the asylum system.

The abuse policy narrative puts forward that the complex problems in Western European asylum systems are caused by human actions. In this way, the abuse policy narrative is able to construct a policy problem that is amenable by policy intervention (Stone 1989). Such an intentional causal mechanism narrative, where the villain is engaging in intentional nefarious action, is often employed in intractable policy debates (McBeth et al. 2012).

1 Shanahan et al. (2017, 177-178) distinguish between three narrative strategies: scope of conflict, causal mechanism, and the devil-angel shift. Stone (1989) furthermore defines four causal mechanisms that can be applied to policy narratives: intentional, inadvertent, accidental, and mechanical.

5

Setting

To accurately represent the setting, we locate asylum policies as a distant and visible policy field that triggers intense public debates, we highlight direct democracy as a central feature of the Swiss policy process and we review the history of asylum reforms in Switzerland.

Soss and Schramm (2007) develop a two-dimensional framework to examine policy fields’ potential for creating mass feedback effects. The policy field’s proximity and visibility to mass publics determines its potential for creating mass feedbacks effects. Policy fields with a high potential of mass feedback effect are more likely to be politicized. Proximity describes the extent to which a policy field affecting the electorate’s lives in immediate, concrete ways (Soss and Schramm 2007). It is assumed that proximate policies are better evaluated by the public without having to rely too heavily on media and elite interpretations, whereas more distanced policies allow political elites to better frame or narrate a policy according to their own specific needs (Hinterleitner 2018). Visibility describes the salience of the policy to the mass public

(Soss and Schram 2007). Highly visible policies are likely to produce stronger feedback effects and thus, these debates are more likely to be staged in public and narrated according to policy narratives.

Asylum policy is a prime example of a distant and visible policy field. Given its potential to create high feedback effects, the public debates in Switzerland around asylum policies are intense and frequent. Over the last thirty years, only a few other policy domains have engendered such regular and heated controversies as asylum policies. Although asylum-seekers usually account for around 10 to 15 percent of immigrants in Switzerland, this category of immigrants has preoccupied politicians and citizens more than any subset of the country’s foreign population in recent years (Bernhard 2012).

Policy narrative are frequently used in Swiss direct democratic campaigns (Schlaufer 2018).

Policy narratives may play the role of activating and reinforcing the previously held opinions of citizens and mobilize them to vote (Kear and Wells, 2014; Schlaufer 2018). The institutions 6

of direct democracy provide political organizations with a powerful opportunity to politicize asylum issues. The issue of asylum has ranked high on the Swiss political agenda over recent years is primarily attributable to the continuous mobilization by mainly Swiss People’s Party

(SVP) that rather effectively tapped into the xenophobic potential among the citizenry

(Skenderovic 2009; Bernhard 2012; D’Amato 2014).

Over the last three decades, Swiss asylum policies have resembled a permanent construction site. Since the introduction of the federal asylum law in 1981, it has been subject to no less than eleven major revisions, including a complete revision in 1999. Many of these reforms of the federal asylum law were challenged by actors from the civil society (such as human right groups, refugee aid organization, charities as well as churches) and sometimes from right-wing dissidents. In addition, the SVP launched two popular initiative about asylum policies.2 In total, there have been six referendums against revisions and modifications of the federal asylum law.

These referendums were clearly defeated at the polls in all six instances3, thereby strengthening the positions of the federal authorities and the right.

When taking a closer look at these asylum reforms, we can observe that Swiss decision-makers have relied on two asylum policy trends (Bernhard and Kaufmann 2018). First, the federal

2 In 1996, the party’s popular initiative obtained an unexpectedly high approval rate of 46.3%.

In 2002, the second popular asylum initiative almost passed as 49.9% of participating voters pronounced themselves in favor of a constitutional amendment called ‘against abuses of the right of asylum’. In both cases, the outcome of the vote was interpreted by the federal authorities and the moderate right parties as a profound discontent of Swiss citizens with available asylum policies.

3 Citizens were called to the polls in 1987, 1994, 1999, 2006, 2013, and 2016. The level of popular support for these reforms ranged from 67.3% in 1987 to 78.5% in 2013.

7

authorities have added more restrictiveness to the asylum legislation by engaging in a continuous tightening. Second, they have frequently enacted streamlining policies, which are designed to increase the efficiency of asylum procedures Streamlining refers to efficiency- enhancing asylum policies that aim to speed up asylum procedures and structurally reorganize competences and responsibilities within national asylum systems.

Characters

Policies and policy target groups are socially constructed: “Policy teaches lessons about the type of groups people belong to, what they deserve from government, and what is expected of them” (Schneider and Ingram 1993, 340). Social constructions are stereotypes about particular groups of people that have been created by politics, culture, or the media. These stereotypes depict the perceptions of citizens and they pervade all aspects of political reality (Pierce et al.,

2014). Policy-making and political behavior can be directed by these stereotypes “without personally endorsing such stereotypes, without feelings of prejudice, and without awareness that such stereotypes could affect one’s judgement and behavior” (Vescio and Weaver 2013, 1; see also Thomann and Rapp 2018).

These constructions are often conflicting between political groups and they subject to political contention (Schneider and Ingram 1993). Asylum seekers are portrayed as either victims or villains, in the terminology of the NPF, or as either dependents or deviants, in the terminology of the social construction of target groups. On the one hand, asylum seekers can be constructed as victims/depends that are in need of protection and that are struggling from strict asylum policies. On the other hand, asylum seekers can be constructed as villains/deviants that abuse generous asylum systems. This social construction of policy target groups depends on how people perceive the deservingness of asylum seekers (Schneider and Ingram 1993, Thomann and Rapp 2018).

8

The abuse policy narrative portrays asylum seekers as economic migrants and bogus refugees who are accused of trying to abuse the country’s generosity and protection. Such a social construction of asylum seekers is prevalent in public debates in Europe. For example, Sales

(2002) shows in political debates about asylum in the United Kingdom that asylum seekers are categorized as ‘bogus’ and ‘genuine’ asylum seekers, with the former being seen as undeserving of sympathy and support. Goodman and Speer (2007) find in an analysis of public debates about asylum that the ways in which asylum seekers are constructed is always a fundamentally politically. Sager and Thomann (2017) reveal that different social constructions of asylum seekers as a policy target group shape the devised asylum policies in Switzerland, mainly because these different social constructions influence the framing of the policy problem.

Furthermore, Crawley and Skleparis (2018) stress how such different social construction of refugees were important in the public debates that heated up during and in the aftermath of the so-called European ‘refugee crisis’4 of 2015.

In direct-democratic campaigns, the portrayal of another type of actor becomes crucial. The homogenously constructed ‘people’ are hyped as the hero, in the NPF language, who put an end to the abuse of the Swiss asylum system. This heroization of the people who solve the

‘asylum problem’ that political elites do not want or cannot solve, attributes a populist characteristic to the abuse policy narrative. Populism perceives societies to be separated into

‘the people’ versus ‘the elite’ and it argues that politics should be an expression of the volonté générale (general will) of the people (Mudde 2004, 543).

4 We put ‘refugee crisis’ in quotation marks because we perceive that it was not just caused by high numbers of asylum seekers, but also by the failure of European governments to formulate viable political solutions to this humanitarian emergency. In this sense, it is rather a crisis of

European asylum institutions than of high numbers of refugees.

9

Plot

The plot establishes the relationships between the characters and situates the characters within the setting. The plot of the abuse policy narrative can be constructed as follow:

The large majority of asylum seekers are economic migrants and bogus refugees. They

abuse the generosity and protection of the Swiss asylum system. They do not only unduly

benefit from generous advantages provided by asylum systems, but they also consume the

resources that should benefit the minority of ‘genuine refugees’. ‘The people’ take back

control of the asylum problem issue that political elites and bureaucrats can not or want

not solve’ by voting for tighter asylum polices in direct-democratic initiatives and

referendums. Thus, ‘the people’ are putting an end to the abuse of the Swiss asylum systems

by the large majority of asylum seekers.

Moral of the Story

The policy solution is tighter asylum policies. The tightening of asylum policies mainly targets two asylum policy fields (see Bernhard and Kaufmann 2018). Frist, the access to the asylum systems for asylum seekers can be made more difficult. States can try to impede the entry of asylum seekers into their territories what prevents that one can lodge an asylum application and enter the refugee status determination procedure. Second, stricter rules for asylum seekers that are already in the asylum procedure can be implemented. Tightening can target the refugee status determination procedure, welfare benefits and the living conditions of asylum seekers. It is argued in the abuse policy narrative that such a tightening would reduce the attractiveness of

Switzerland as an asylum destination, especially for those asylum seekers that solely chose

Switzerland as a destination because of its generosity. Thus, the moral of abuse policy narrative is that tighter asylum policies would solve the asylum problem because they would make it harder for asylum seeker to abuse the national asylum system.

10

Hypotheses

We formulate one hypothesis with regard to the policy orientation of asylum policy reforms and one hypothesis about the political ideology of the political elite actors and citizens.

With regard to the policy orientation, we expect that political elite actors are more likely to convey the asylum abuse policy narrative in the context of tightening asylum reforms than when streamlining is at stake. Accordingly, the abuse policy narrative is more important for citizen’s opinion formation in tightening asylum reforms than in streamlining asylum reforms.

Tightening asylum policies try to either limit access to asylum applications or attempt to reduce the attractiveness of an asylum destination country, whereas streamlining asylum policies try to enhance the efficiency of asylum systems by speeding up asylum procedures and structurally reorganize competences and responsibilities within national asylum systems (Bernhard and

Kaufmann 2018). As we have outlined, the moral of the abuse policy narrative is that tighter asylum policies would solve the ‘asylum problem’. Streamlining asylum policies have a different policy rationale. The objectives of acceleration and cost reduction, which are fundamental in streamlining reforms, are rather in line with a pragmatic problem-solving approach and it does not fundamentally question the liberal asylum regime (Bernhard and

Kaufmann 2018).

H1: The more a reform includes tightening policies, the more political elite actors rely

on the abuse policy narrative and the more important is the abuse policy narrative for

citizen’s opinion formation.

With regard to the political ideology, we expect that a right political ideology makes political elite actors and citizens more aligned to the asylum abuse policy narrative. Conflicts over asylum policies in Western Europe are structured along the left–right axis as they are rooted in the new cultural conflict dimension (Kriesi et al. 2008; Akkerman 2015). This conflict 11

dimension has become increasingly politicised as a result of mobilizations by the new left and the Greens as well as by the radical right (Bornschier 2010). Right political actors have continuously mobilized their voters around a tightening of asylum policies, while the left and its civil society allies have kept trying to maintain the status quo or strengthen the rights of asylum seekers. Actors from moderate parties take generally an intermediate stance (Kriesi et al. 2008) that can be also observed in immigration issues (Akkerman 2015). This conflict structure is also observable in Switzerland: While the radical right blamed the federal authorities for not doing enough to prevent abuses of the asylum system, political actors from the left and NGOs accused the right of violating humanitarian principles (Skenderovic 2007,

Steiner 2000). The specific narrative of the radical right has revolved around asylum seekers as economic and bogus refugees who are accused of trying to abuse the country’s generosity

(Skenderovic 2009).

H2: The more to the right political ideology of political elite actors and citizens, the more

political elite actors convey the abuse policy narrative and the more important is the

abuse policy narrative for citizen’s opinion formation.

Research Design

The empirical analysis investigates the use and the effect of the abuse policy narrative in the contexts of the three most recent asylum referendum campaigns that have taken place at the

Swiss federal level. The asylum referendums in Switzerland are highly politicized. Since citizens had the final say on the policy packages, political elites had a strong incentive to go public during the campaign that preceded the vote by relying on policy narratives. It is under such circumstances that political elite actors of various political stripes make use of political narratives and that citizens are confronted with these policy narratives in their opinion formation. Due to this politicization, the selected asylum referendums present an opportunity 12

for examining the policy narratives of a large population of political actors, or so-called elite actors, as well as to study the influence of this narrative on citizens’ opinion formation. In addition, these three referendums vary greatly in terms of their policy orientation. Hence, we can also compare the impact of the policy orientation on the use and the effect of the abuse policy narrative.

The cases under scrutiny

In the following, we briefly summarize the main contents of the three referendums under investigation. The revision citizens adopted in 2006 was characterized by some highly controversial tightening provisions. These included the prohibition of social assistance for asylum seekers whose requests had been rejected, more restrictive rules for asylum-seekers without proper identification, more drastic coercive measures, and more restrictive asylum procedures at airports. The bill also stipulated that requests from asylum-seekers that were already treated by so-called secure third states would not be proceeded by Swiss authorities.

The urgent modifications of the asylum law that Swiss citizens adopted in 2013 were composed of four tightening and four streamlining provisions. With respect to the tightening policies, the provisions include the ban of the possibility to apply for asylum at Swiss embassies, the denial of refugee status for deserters, the creation of special centres housing so-called ‘troublemakers’

(i.e. asylum seekers who refuse to cooperate with the authorities), and the reduction of appeal periods for rejected asylum requests. With respect to the streamlining policies, the provisions include the testing of the then planned huge structural reforms in a reception centre, greater powers of the federal authorities over the cantons (i.e. the 26 member states) to cope with the challenge posed by volatile numbers of asylum seekers, as well as two measures that provides the cantons with higher federal subsidies (i.e. flat-rate subsidies for covering additional security costs and financial contributions to employment programs).

13

The referendum of 2016 was fully geared to a streamlining reform that the Swiss authorities had modelled on the Dutch example. The goal of the asylum provision was to enhance procedures in order that the refugee status determination takes no longer than a year, as opposed to around two years before the reform. To speed up refugee determination procedures, the provision wanted to create new reception centres ran by the Confederation that should deal with likely negative asylum applications, asylum seekers are only distrusted to the cantons for complex and likely positive cases that require further evaluations of the asylum claims, and asylum-seekers are given the right to benefit from free legal advice.

Data collection and methods

We mainly rely on an original data set that we complied by conducting 108 face-to-face interviews with campaign managers of those organizations that actively participated in the referendum campaigns of 2006, 2013, and 2016 (see Table A1 in the appendix). We selected these actors on the basis of various sources such as the parliamentary debates, the campaign for the collection of signatures, voting recommendations, ore the press. In addition, cross- referencing was used with the interviewees to complete the set of organizations. Given this pragmatic procedure, we feel confident to have included the most important organizations. We interviewed 46 political actors for the 2006 referendum, and respectively 31 for the two other cases. With each campaign manager, we performed an ex ante interview by relying on a structured questionnaire that contained about 100 closed-ended or open-ended questions. The interviews usually lasted between 35 minutes and an hour. We analyze these interview data with descriptive and multivariate methods.

14

In order to assess, the effect of the abuse policy narrative on citizens’ opinion formatoin, we make use of the VOX post-vote surveys5 of all three direct-democratic votes (Longchamp et al.

2006; Longchamp et al. 2013; Longchamp et al. 2016). These surveys enable us to study the difference in the importance of the abuse policy narrative for citizens’ opinion formation between the three direct-democratic campaigns under scrutiny.

Operationalization

We measure the salience of the abuse policy narrative, our dependent variable, by relying on a series of questions about key narratives or messages we believed could be important in the three referendums of interest. In the framework of ex-ante interviews, campaign managers were confronted with a list of 12 key narratives and messages. For each of the selected referendum, one of the 12 key narratives or messages refers to abuse policy narrative.6

This allows us to measure the relative salience political organizations devoted to the abuse policy narrative. We asked the various campaign managers to classify these messages according to the importance they would attach to them during their campaign communication. We first invited the respondents to choose the three most important messages. Among these three messages, we then asked them to select the most important one. As to the nine remaining messages, campaign managers finally had to mention the three least important ones. Based on

5 These VOX surveys were conducted during the two or three weeks following all Swiss direct- democratic votes from 1977 to 2016 by standardized survey via telephone interviews.

6 We relied on the following wording: ‘Abuses in the domain of asylum have to be stopped’.

The wording of the concrete questions had to be parsimonious given that we presented the interviewees 12 key narratives and messages as well as over 100 questions in total. In the context of these direct democratic campaigns, it should have been clear that this wording refers to the more extensive abuse policy narrative that was frequently put forward by political actors.

15

the answers we obtained, we constructed a scale that ranges from 0 to 3 by applying the following scheme:

– 3 for the single most important message,

– 2 for the following two important messages,

– 1 for the six moderately important messages, and

– 0 for the three least important messages.

Our first independent variable are the three different ballot measures (we set the tightening reform of 2006 as reference group) to test the first hypothesis. To test the second hypothesis, we make use of the left-right placement of the selected organizations as the second independent variable. As it is usual for survey questionnaires, this scale ranges from 0 (completely left) to

10 (completely right).

In addition, we control for the influence of three variables: support of the abuse policy narrative of the selected organizations, actor types, and language region. With regard to the control

‘support for the abuse policy narrative’, we rely on a question available from the ex-ante interviews with campaign managers. The respondents were asked to indicate the positions of their respective organizations about the abuse message by using a five point-scale that ranged from (1) ‘do not agree at all’ to (5) ‘totally agree’. It can be assumed that the more a political organizations support the very content of the abuse policy narrative, the more a political organization uses this policy narrative. Regarding actor types, we draw a distinction between political parties, interest groups, committees, and state actors (see Table A1 in the appendix). It may be the case that these political actors vary in their propensity to convey the abuse policy narrative in direct democratic campaigns. The modal category ‘interest groups’ is defined as reference category. We control for the language region because elite actors of the French language regions tend to hold a more liberal position on immigration-related issues than elite actors in the rest of the country (Bernhard 2012, 47). We construct a dichotomous variable by

16

coding the organisations based in the French-speaking part of Switzerland (n=16) as ‘1’, while the remaining ones take the value of ‘0’.

Results

To account for the conveyance of the abuse policy narrative by pollical elite actors, we first descriptively examine the salience of the abuse message before we test our hypotheses on the basis of multivariate analysis. In a next analytical step, we scrutinize the effect of the abuse policy narrative on citizens’ opinion formation by analyzing post vote surveys.

The conveyance by political elite actors

The salience of the abuse message, our dependent variable, varied in the three different ballot measures. The average score among interviewed political actors turns out to be highest for the

2006 campaign. Our salience measure amounts to 1.35 for the 2006 referendum that was dominated by tightening provisions. With an average of 1.16, the campaigners’ emphasis of the abuse message turned out to be somewhat inferior in the context of the 2013 urgent modifications of the asylum law, which we consider a balanced case in terms of policy orientation. The lowest level is obtained for the streamlining reform of 2016. For this campaign, the average score reaches only 0.81. Hence, these scores are in line with the first hypothesis.

To obtain a first impression of the empirical validity of the second hypothesis of interest, we ran a Spearman’s rank correlation test, given that both the left-right placement and the abuse salience indicator are discrete ordinal variables. The coefficient (rho=0.54) turns out to be positively significant at the 0.1% error level. It thus appears that the salience of the abuse narrative message increases the more to the right a given political actor places itself. This lends support to the second hypothesis.

We now test whether these observed patterns are confirmed when relying on a multivariate analysis. The results of the ordered probit model are shown in Table 1. In accordance with the

17

first hypothesis, it turns out that political organizations made less use of the abuse policy narrative during the campaign of the 2016 streamlining reform than on the tightening reform.

Regarding the second hypothesis, the salience of the abuse message increases the more to the right a political actor positions itself. Hence, both hypotheses can be also confirmed according to our multivariate analysis.

18

Table 1: Ordered probit model explaining the salience of the abuse policy narrative Model (1)

Asylum 2013 -0.017 (-0.06)

Asylum 2016 -0.830** (-2.64)

Left-right political ideology 0.220*** (scale 0-10) (3.31)

*** Support for abuse policy narrative 1.071 (4.24)

Parties 0.251 (0.88)

Committees -0.664 (-1.26)

State actors -0.176 (-0.39)

French-speaking part -0.161 (-0.44)

Cut 1 -0.302 (-0.87)

Cut 2 2.149*** (4.93)

Cut 3 2.995*** (6.14) N 104 Pseudo R2 0.318 * p<0.05 ** p<0.01 *** p<0.001; z-values in brackets. Reference categories: Asylum 2006 (for ballot measures) and interest groups (for actor types)

19

As to the control variables, levels of salience of the abuse message are found to be very closely related to the actor positioning on this message. Indeed, the reliance on the abuse message strongly increases with the level of agreement with it. By contrast, we do not report any significant differences across actor types. The same holds true with respect to language regions, as organisations from the French-speaking part do not put less emphasis on the abuse message than the remaining organizations. This means that the abuse policy narrative is commonly used by actors from the political right whether they are parties, interest groups, or committees and whether they are based in the German or the French part of Switzerland.

The effect on citizens’ opinion formation

We make use the VOX post-vote surveys to assess and compare the effect of the abuse policy narrative for citizens’ opinion formation in the three direct-democratic campaigns under scrutiny. The VOX post vote surveys ask a representative sample of eligible voters various questions regarding the last vote. We make use of a standard open question were citizens could state the main reason for their voting behaviour. We coded whether the main reason for the vote

20

was related to the asylum abuse policy narrative.7 We also utilize a question that asks citizens to place themselves on a political ideology specturum between ‘radcial left’ and ‘radical right’.8

Table 2 reports whether the main reasons for citizens opinion formation was related to the asylum abuse policy narrative segmented by political ideology of citizens. Reasons related to the abuse policy narrative were crucial in the 2006 referendum as more than one quarter of citizens report such reasons as crucial for their opinion formation. This is especially striking compared to the other two votes where the asylum abuse policy narrative was of marginal importance, but is was more important in the balanced reform of 2013 than in the streamling reform of 2016. Thus, we can support hypothesis 1 because it is more likely that the asylum

7 The open answers are issue specific. To categorize the open answers, we make use of inductively build categories made by VOX post-vote survey analysts. For the 2006 vote, we included the open answers that fall in three following inductively build categories: (1) “abuse”

(general), (2) there is too much abuse”, “we have to fight abuse”, “stop abuse”, “curb the abuse”,

(3) “stricter handling of ‘bogus’ asylum seekers”, “’real’ asylum seekers are not put in a worse position”, and “more control”. For the 2013 vote, we included open answers that fall in three inductively build categories: (1) “only ‘real’ asylum seekers should be granted asylum (no

‘economic’ asylum seekers etc.) (2) “against the abuse of asylum/there is too much abuse”, (3)

“there are too much ‘economic’ asylum seekers”. For the 2016 vote, we included the open answers that fall in two inductively build categories: (1) “short proceeding deter ‘economic’ asylum seekers” and (2) “there are too much ‘economic’ asylum seekers”.

8 We rely on the ideological self-placement of citizens that ranges from 0 (completely left) to

10 (completely right). We assigned the values of 0 and 2 to the radical left, those from 3 and 4 to the moderate left, 5 to the center, 6 and 7 to the moderate right, and 8 to 10 to the radical right.

21

abuse policy narrative is important for citizen’s opinion formation in refomrs that include tightening policies

Table 2: Effect of the asylum abuse policy narrative on citizens’ opinion formation

2006 2013 2016 Reasons related to the abuse 26.25 % 5.23 % 1.68 % policy narrative (121) (39) (16) Political ideology Radical right 36 % 6.19 % 0 % (18) (6) (0) Moderate right 35.96 % 5.47 % 1.01 % (32) (7) (2) Center 33.57 % 8.58 % 3.14 % (48) (20) (7) Moderate left 8.82 % 0.65 % 1.81 % (9) (1) (3) Radical left 9.26 % 2.11 % 2.72 % (5) (2) (4) No answer/don’t know 24.32 7.69 % 0 % (9) (3) (0) Total respondents (N) 461 745 953

With regard to the second hypothesis, we can note that center, moderate right and radical right citizens reported more frequently that the asylum abuse policy narrative was the most important reason in for their vote. This pattern is best visible in the 2006 as more than one third of center to radical right citizens were primarily influence by the asylum abuse policy narrative. The pattern also holds in the 2013 but we have to be cautious in the interpretation given the low number of respondents, whereas there is no pattern in the 2016 referendum and the number of responds is even lower. We can not say that the importance of the asylum abuse policy narrative increases the more to the right citizens placed their ideological positioning on the left-right scale and thus we can not support hypothesis 2. However, we can state that the asylum abuse policy narrative is rather equally important for citizens opinion formation for citizens from the centre to the radical right and that this is most visible when tightening policies are at stake in a referendum.

22

Conclusion

We study the use as well as the effect of the abuse policy narrative in three direct democratic campaigns on asylum policy in Switzerland. The abuse policy narrative is conceptualized by the structure presented in the Policy Narrative Framework. We combine a mesolevel and microlevel analysis by studying which political elite actors convey the abuse policy narrative and when they convey it as well as by testing the effect of the abuse policy narrative on citizens’ opinion formation. We conducted 108 structured interviews with campaign managers in advance of these asylum referendums and we analyzed these interviews with an ordered probit regression analysis. In addition, we utilized post vote surveys to examine the effects of the asylum abuse policy narrative on citizens’ opinion formation.

We find that the more to the right political elite actors are, the more they convey the asylum abuse policy narrative but only in referendums about the tightening of asylum policies. With regard to citizens’ opinion formation, the asylum abuse policy narrative is crucial in tightening reforms for citizens from the center, the moderate right and was well the radical right in these tightening reforms. The findings in this article corroborates conventional wisdom that actors from the right rely on the abuse policy narrative. However, it adds more nuance to it by showing that the abuse policy narrative is mostly conveyed when tightening policies are at stake. The results are similar for the conveyance of the abuse policy narrative by political elite actors as well as for the effect of the abuse policy narrative on citizens’ opinion formation. Citizens from the center to the radical right are more influenced by the asylum abuse policy narrative in votes about tightening policies. These findings also support our construction of the abuse policy narrative because we conceive tighter asylum policies as the solution to the constructed problem in the structure of the asylum policy narrative. Thus, the tightening of asylum polices seems to intrinsic to the very design of the abuse policy narrative, whereas the logic of streamling policies seems not to fit with the abuse policy narrative.

23

Beyond these concrete findings, these article offers a systematic conceptualization of the popular accusation that asylum seekers try to abuse a country’s generosity and protection as a policy narrative that is prevalent in the popular discourse in many asylum destination countries

(e.g. Schuster 2004; Hänggli and Kriesi 2010; Zagor 2015). We could show that these narratives is not only conveyed by political elite actor form the right but has an influence on citizens’ opinion formation on asylum policy. This asylum abuse policy narrative is not only resonating with right-wing citizens, but as well with neutral and moderate right citizens, at least when tightening asylum policies are at stake.

We expect that this conceptualized abuse policy narrative is of use to study public debates about asylum policies beyond referendum campaigns in Switzerland and that the results offers clues in which context which actors make use of the abuse policy narrative. We expect that the abuse policy narratives in other contexts is also not employed randomly but it is strategically activated and conveyed by actors from the political right when tightening policies are at stake. However, because these debates about Swiss asylum policies were held in the context of direct democratic campaigns, the abuse policy narratives may be used more intensively and more frequently than in other types of public debates.

We were able to show that the accusation that a large proportion of asylum seeker are economic migrants and bogus refugees trying to abuse the country’s generosity and protection is a strategically used political narratives that this abuse policy narrative is convincing for large proportion of citizens. Given that the total protection rate of asylum seekers is globally about

50%9 (UNHCR 2018, 42-45), the claim that there is a widespread abuse of national asylum

9 The protection rate of can be calculated as the proportion of asylum-seekers that were granted the Convention status or another form of protection of the total number of substantive decisions. (UNHCR 2018, 42).

24

systems is overstated at best. The reason for the popularity of the abuse policy narrative is therefore not the validity of the narrative but seems to be its persuasiveness. This persuasiveness may be explained by the narrative’s activation of the distinction between refugees ‘worthy’ of protection and ‘bogus’ (economic) migrants that is enshrined in asylum policy practices of many state actors (De Genova 2002; Darling 2014; Skleparis and Crawley 2018). Against this background, this article importantly deconstructs this dangerous narrative of asylum seekers because it outlines its very construction and utilization by actors from the political right. The narrative is dangerous because it convinces citizens well into the middle of the political spectrum.

25

References

Akkerman, Tjitske. 2015. “Immigration Policy and Electoral Competition: A Fine-grained

Analysis of Party Positions over the Past Two Decades.” Party Politics 21 (1): 54–67.

Bernhard, L. 2012. Campaign Strategy in Direct Democracy. Basingstoke: Palgrave

Macmillan.

Bernhard, L., und D. Kaufmann. 2018. Coping with the Asylum challenge: Tightening and

Streamlining Policies in Western Europe. Journal of Ethnic and Migration Studies 44(15):

2506–2523.

Bloch, Alice. 2000. “A new era or more of the same? Asylum policy in the UK.” Journal of

Refugee Studies 13(1): 29–42.

Bornschier, Simon. 2010. Cleavage Politics and the Populist Right: The New Cultural Conflict

in Western Europe. Philadelphia, PA: Temple University Press

Crawley, H. and D. Skleparis. 2018. Refugees, Migrants, Neither, Both: Categorical Fetishism

and the Politics of Bounding in Europe’s ‘Migration Crisis’. Journal of Ethnic and Migration

Studies 44 (1): 48–64

Crow, D. 2012. “The Narrative Policy Framework: Broadening the Framework for Increased

Relevance”, Paper presented at the 2012 Midwest Political Science Association meetings,

Chicago, April 12-15.

Czaika, M., and H. De Haas. 2013. “The Effectiveness of Immigration Policies.” Population

and Development Review 39(3): 487-508.

D’Amato, G. 2014. “Switzerland.” In: Hollifield, J. und P. M. Martin, Hrsg. Controlling

Immigration: A Global Perspective (pp. 308-322). Stanford: Stanford University Press.

Darling, Jonathan. 2014. Asylum and the post‐political: domopolitics, depoliticisation and acts

of citizenship. Antipode 46(1): 72–91. 26

De Genova, Nicolas. 2002. Migrant “illegality” and deportability in everyday life. Annual

Review of Anthropology 31:419–447.

Goodman, Simon and Susan A. Speer. 2007. Category Use in the Construction of Asylum

Seekers, Critical Discourse Studies 4(2): 165–185.

Hänggli, Regula and Hanspeter Kriesi. 2010. “Political Framing Strategies and Their Impact on

Media Framing in a Swiss Direct-Democratic Campaign” Political Communication 27(2):

141-157.

Hatton, T. 2009. “The Rise and Fall of Asylum: What Happened and Why?” The Economic

Journal 119(535): 183-213.

Hatton, T. 2012. Seeking Asylum: Trends and Policies in the OECD. London: Centre for

Economic Policy Research (CEPR).

Hinterleitner, Markus. 2018. Policy failures, blame games and changes to policy practice.

Journal of Public Policy 38.2 (2018): 221–242.

Inderbitzin, I. 2002. The Foreignisation Process in Switzerland: The Swiss and their Ausländer.

Dissertation, Monash University.

Jennings, Will. 2010. “Bureaucratic performance and control in British politics: asylum policy

1994–2007.” The British Journal of Politics and International Relations 12(4): 539–568.

Jones, Michael D., and Mark K. McBeth. 2010. “A Narrative Policy Framework: Clear Enough

to be Wrong?” Policy Studies Journal 38 (2): 329–53.

Kear, Andrew R., and Dominic D. Wells. 2014. “Coalitions Are People: Policy Narratives and

the Defeat of Ohio Senate Bill 5.” In The Science of Stories: Applications of the Narrative

Policy Framework in PublicPolicy Analysis, ed. Michael D. Jones, Elizabeth A. Shanahan,

and Mark K. McBeth. Basingstoke, UK:Palgrave Macmillan, 157–84.

27

Kriesi, H., E. Grande, R. Lachat, M. Dolezal, S. Bornschier, and T. Frei. 2008. West European

Politics in the Age of . Cambridge: Cambridge University Press.

Kriesi, H., and A. H. Trechsel. 2008. The Politics of Switzerland. Cambridge: Cambridge

University Press.

Longchamp, C., Linder, W., Hardmeier, S., Sciarini, P., Golder, L., Milic T. 2006. Vox 91 -

Votation du 24.09.2006 [Dataset]. https://doi.org/10.23662/FORS-DS-729-1

Longchamp, C., Vatter, A., Widmer, T., Sciarini, P. Imfeld, M., Tschöpe, S. 2013. Vox 111 -

Votation du 09.06.2013 [Dataset]. https://doi.org/10.23662/FORS-DS-707-1

Longchamp, C., Vatter, A., Widmer, T., Sciarini, P., Imfeld, M., Tschöpe, S. 2016. Vox 121 -

Votation du 05.06.2016 [Dataset]. https://doi.org/10.23662/FORS-DS-825-1

McBeth, Mark K., Elizabeth A. Shanahan, Molly Anderson, and Barabara Rose. 2012. “Policy

story or gory story? Narrative policy framework analysis of Buffalo field campaign’s

YouTube videos.” Policy & Internet 4: 159–183.

McBeth, Mark K., Michael D. Jones, and Elizabeth A. Shanahan. 2014. “The Narrative Policy

Frame-work.” In Theories of the Policy Process, ed. Christopher M. Weible, and Paul A

Sabatier. Boulder, CO: Westview Press, 225–6.

McBeth, Mark K. and Donna L. Lybecker. 2018. “The Narrative Policy Framework, Agendas,

and Sanctuary Cities: The Construction of a Public Problem.” Policy Studies Journal 46(4):

868–893.

Mudde, Cas. 2004. “The populist zeitgeist.” Government and Opposition 39(4): 541–563.

Pierce, Jonathan J., Saba Siddiki, Michael D. Jones, Kristin Schumacher, Andrew Pattison, and

Holly Peterson. 2014. “Social Construction and Policy Design: A Review of Past

Applications.” Policy Studies Journal 42 (1): 1–29.

Sales, R. (2002). The deserving and the undeserving. Critical Social Policy, 22, 456–478. 28

Sager, Fritz, and Eva Thomann. 2017. “Multiple streams in member state implementation:

politics, problem construction and policy paths in Swiss asylum policy.” Journal of Public

Policy 37(3): 287-314.

Schlaufer, Caroline. 2018. “The narrative uses of evidence.” Policy Studies Journal 46(1): 90-

118.

Schneider, Anne, and Helen Ingram. 1993 “Social construction of target populations:

Implications for politics and policy.” American Political Science Review 87.2 (1993): 334-

347.

Schuster, Liza. 2004. The use and abuse of political asylum in Britain and Germany. Routledge.

Schuster, Liza, and John Solomos. 2004. “Race, immigration and asylum: New Labour’s

agenda and its consequences.” Ethnicities 4(2): 267–300.

Shanahan, E. A., Jones, M. D., McBeth, M. (2011). “Policy narratives and policy processes.”

Policy Studies Journal 39 (3), 535–561.

Shanahan, E. A., Jones, M. D., McBeth, M. K., & Radaelli, C. M. (2017). The narrative policy

framework. In C. M. Weible & P. A. Sabatier (Eds.), Theories of the Policy Process (pp.

173–213). New York: Westview Press.

Skenderovic, D. 2009. The Radical Right in Switzerland: Continuity and Change, 1945-2000.

New York: Berghahn Books.

Soss, Joe, and Sanford F. Schram. 2007. “A public transformed? Welfare reform as policy

feedback.” American Political Science Review 101(1): 111–127.

Steiner, N. 2000. Arguing about Asylum: The Complexity of Refugee Debates in Europe. New

York: St. Martin’s Press.

Stone, Deborah A. 1989. “Causal stories and the formation of policy agendas.” Political Science

Quarterly 104(2): 281–300. 29

Vescio, Theresa, and Kevin Weaver. 2013. “Prejudice and Stereotyping.” Oxford

Bibliographies Online: https://doi.org/10.1093/obo/9780199828340-0097. Accessed xxx.

Thomann, Eva, and Carolin Rapp. (2018). “Who Deserves Solidarity? Unequal Treatment of

Immigrants in Swiss Welfare Policy Delivery.” Policy Studies Journal 46(3): 531–552.

UNHCR. 2018. Global Trends: Forced Displacement in 2017. Online:

https://www.unhcr.org/5b27be547.pdf

Zagor, Matthew. 2015. “The struggle of autonomy and authenticity: framing the savage

refugee.” Social Identities, 21(4): 373–394.

30

Appendix

Table A1: List of organizations according to political groups Political Actor type 2006 2013 2016 orientation Radical left Political party Greens Young Socialists Basel Alternatives Young Socialists Young Socialists Communists State actor Interest group/ Solidarity without Swiss Worker’s Aid Nongovernmental frontiers Group for a organizations Switzerland without an army Committee Comedia Living Together Solidarity Solidarity Moderate left Political party CJPLD Evangelicals Evangelicals FIMM Greens Greens GGK Young Greens Social Democrats Politakt Social Democrats Social Democrats State actor Directors of Cantonal Social Services Interest group/ AGORA Evangelical Social Caritas Nongovernmental Evangelical church of Center HEKS organizations Democratic Lawyers Swiss Aid for HEKS Solidarity with Refugees Collective of Irregulars frontiers Federation of Swiss Aid for Refugees Unia Evangelical Churches Swiss Federation of Swiss Worker’s Aid Trade Unions Solidarity without Unia frontiers Unia Committee Committee 2xNo Stop exclusion Appeal for the Committee of cultural maintenance of the Actors right of asylum Stop exclusion Stop exclusion Center Political party Christian Democrats Christian Democrats Green Liberals (federal level as well as Green Liberals Operation Libero VD and GE) Evangelicals State actor Federal Office for Federal Office for Federal Justice and Migration Migration Police Department Federal Justice and State Secretariat for Police Department Migration Interest group/ Amnesty international Amnesty international Amnesty international Nongovernmental Swiss Bishop’s AGORA Justice and Peace organizations Conference Caritas Swiss City Association Federation of HEKS Evangelical Churches Justice and Peace VJSF Federation of Evangelical Churches Committee Coalition in favor of a Humanitarian Switzerland Right-wing Committee against Asylum Law Moderate Political party Radical Democrats Conservative Conservative right (federal level and GE) Democrats Democrats Liberals Radical Liberals Christian Democrats Geneva Citizen Radical Liberals Movement Swiss Democrats State actor Federal Justice and Cantonal Justice and Cantonal Justice and Police Department Police Directors Police Directors

31

Interest group/ Young4fun.ch Young4fun.ch Swiss Homeowner Nongovernmental Association organizations Committee Radical right Political party Evangelical Young Swiss People’s Swiss People’s Party Fundamentalists Party Party Swiss People’s Party Swiss People’s Party

State actor Interest group/ Action for an Pro Libertat Nongovernmental Independent and organizations Neutral Switzerland Swiss Small Business Association Committee PIKOM Security for All

32