• 1 Introduction

  When citizens accept initiatives in the framework of a direct-democratic ballot, there is often great jubilation among their originators. This is all too understandable, since such victories usually are obtained over large parts of established political elites and aim at substantial policy change. However, the demands of accepted initiatives do not automatically take effect with the people’s verdict at the polls. After their passage, such initiatives find themselves in fact at the very beginning of a long, arduous and often frustrating process for their supporters. As Gerber et al. (2001, p. 5) have forcefully emphasised, initiatives never implement nor enforce themselves. For accepted initiatives to affect public policies, political actors in charge of the implementation – typically legislators, but sometimes also members of government and civil servants – have to comply with their core provisions. Yet a major challenge arises from powerful post-passage opposition by precisely those actors that are responsible for implementation decisions (Lupia & Matsusaka, 2004). For instance, many initiatives contain policies that MPs prevented from passing via legislative channels at an earlier point in time (Gerber et al., 2004). Asking lawmakers to perform their duty by faithfully implementing initiatives they dislike may indeed require them to force themselves to do so.
  In line with this leadership challenge, the few empirical analyses that have studied policy implementation in direct democracy thus far have found that numerous initiatives are modified, abandoned or ignored altogether (see Lupia & Matsusaka, 2004 for an overview). Suffice it to say that the actors in charge of implementation expose themselves to harsh criticism from the supporters’ side in such instances. Yet the accusation directed at elites to flout the will of the people has to be taken very seriously. Indeed, a main aim of direct democracy in general and initiatives in particular is to enable marginalised groups to change policies according to their wishes. Non-compliance with accepted initiatives is concerning, as large parts of the citizenry may have the feeling that they are not able to influence the political process and to make their voices heard. This poses a serious democratic challenge. If elites do not faithfully implement policies accepted by citizens, they may not only nurture populist attitudes but also widespread distrust, thereby ultimately calling into question the legitimacy of existing democracies. Therefore, addressing and understanding the reasons for non-compliance of initiatives can be considered important from a normative perspective.
  Despite the relevance of this topic, it is striking that political scientists lack deep knowledge about the determinants of (non-)compliance in direct democracy. Indeed, the state of the art is scant as to the number of publications and restricted in terms of context, given that scholars have limited themselves to the experience made in U.S. states. In addition, there has been an astonishing disregard in recent years. Hence, political scientists may have overlooked some important topical developments as far as lawmakers’ responses to the passage of initiatives are concerned.
  This article seeks to examine the policy implementation literature in direct democracy by focusing on the challenging role played by the supranational level, a neglected factor in the field. Indeed, political scientists have not explicitly addressed the fact that some popular demands that are expressed in winning initiatives may conflict with international law. Hence, such provisions may be difficult to be complied with at the domestic level. More specifically, I argue that legislators (and other actors in charge of implementation) face a dilemma between responsiveness (i.e., respecting the people’s will) and responsibility (i.e., complying with a country’s international obligations). This article examines the provisions of accepted initiatives that conflict with international law provisions legislators are more or less likely to comply with. By focusing on the Swiss context, this article draws a distinction between mandatory and non-mandatory parts of international law (see Section 3). It is hypothesised that legislators will not implement provisions of accepted popular initiatives that conflict with mandatory international law, while they will be more open to comply with provisions that conflict with non-mandatory international law.
  The empirical part of this article is devoted to a case study of the Swiss deportation initiative, a proposition launched by the Swiss People’s Party (SVP) in 2007 that demanded criminal foreigners to be expelled from the country for some offences. The selected initiative clashed with various principles of international law. After its passage in 2010, the implementation of the deportation initiative posed a serious challenge to Parliament, which was in charge of translating the new constitutional provisions into federal legislation. As will be shown, the two chambers of Parliament ultimately decided in 2015 to square the circle between responsiveness and responsibility by enacting a decisive tightening of penal legislation that nevertheless conformed with non-mandatory and above all mandatory provisions of international law. This result proves to be basically in line with the hypothesis. In this case, MPs therefore relied as much as possible on responsiveness and as little as necessary on responsibility.
  The remainder of this article is organised as follows. Section 2 is devoted to the state of the art in the field of policy implementation in direct democracy by focusing on the key factors of (non-)compliance. Subsequently, the restrictions posed by international law are outlined in Section 3. After some general considerations, the Swiss practice in dealing with this challenge is addressed by introducing the crucial distinction between mandatory and non-mandatory norms of international law. Section 4 outlines the selection of the Swiss deportation initiative. Section 5 thereafter presents the implementation of this case. While special attention is given to the bargaining process in Parliament, the pre-parliamentary phase will be also scrutinised, given that it pre-structured the implementation decisions taken by legislators. Finally, Section 6 reviews the major findings of this article and discusses some implications.

• 2 State of the Art

  Despite the normative importance of policy implementation in direct democracy, the political science literature proves to be scant on this topic. To the extent that such studies exist, they focus on U.S. states, especially California. What is more, there is an obvious lack of recent contributions. Taken together, the work by political scientists has nevertheless identified some factors that are key in explaining the likelihood of (non-)compliance with winning initiatives.
  To date, scholars have understandably put much emphasis on the role played by leadership. More specifically, they have focused on the preferences of those actors in charge of implementation (Bali, 2003; Ferraiolo, 2007; Gerber et al., 2001, 2004). In the case of California, Gerber et al. (2001, p. 6) have stated that “if a legislative majority and the governor are united in their opposition to an initiative, then full compliance is extraordinarily unlikely”. Yet this is often the case in any political system with direct-democratic provisions, given that initiatives typically propose policy changes lawmakers dislike. Indeed, proponents would have not taken the costly path of launching an initiative to obtain their desired policy otherwise (Gerber, 1999).
  However, under some circumstances, actors whose support is necessary for compliance can be expected to be less inclined to tinker with a voter-approved initiative. The formal models developed by Gerber et al. (2001, 2004) posit that the level of compliance increases when lawmakers face high sanctions. In line with these general expectations, Bali 2003 and Ferraiolo 2007 have found that the extent of compliance positively depends on the level of popular support for an initiative.
  Another crucial moderating aspect refers to characteristics of the ballot measures themselves. In this respect, the precision of initiatives emerges as the single most important factor from previous work. Accordingly, scholars agree that non-compliance is less likely to occur when winning initiatives provide clear instructions to those responsible for implementation (Ferraiolo, 2007; Gerber et al., 2001, 2004). By contrast, lawmakers’ ability to resist the popular mandate is greatest when the terms are vague in nature.
  Finally, political science scholars have studied the role of three types of institutions when it comes to policy compliance.¹x A reviewer of this article convincingly pointed to an additional crucial institutional characteristic from a theoretical perspective by drawing the distinction between legally binding and advisory direct-democratic votes (Setälä, 1999, p. 338). The former can be reasonably expected to be more strictly implemented than the latter. However, advisory votes may in fact also provide strong guidance to decision-makers (Jäske & Setälä, 2019, p. 91). First, empirical work from the United States suggests that a legislature’s leeway in the stage of implementation can vary considerably across jurisdictions (Ferraiolo, 2007). Whereas California prohibits legislators from revising, rejecting or otherwise altering laws passed by initiative, the room of manoeuvre is much larger in other states. Second, rulings by courts may have a detrimental effect, as they can lead to prohibitively high costs of compliance or prevent it altogether (Gerber et al., 2001, 2004). To be sure, legal studies suggest that numerous ballot measures have posed serious constitutional problems in the United States (Miller, 2009; Schacter, 1995). Third, institutional veto players have to be taken into account. Accordingly, a growing number of actors required to implement and enforce initiatives can generally be expected to favour non-compliance (Gerber et al., 2001, 2004).
  As will be developed in the next section, this article aims at advancing the existing literature by considering an institutional factor in the supranational domain. More specifically, it will address the constraining role played by international law when it comes to policy implementation of initiatives in the domestic level.

• 3 The Dilemma Posed by International Law

  Popular sovereignty and the rule of law are two main principles of liberal democracies. While these principles can be basically considered to be mutually dependent, their relationship does not come without tensions. Given that this topic is particularly salient in direct democracy (Marxer & Pállinger, 2007; Miller 2009), it is striking that the potentially tense relationship between popular sovereignty and the rule of law has not been discussed in detail so far in academia (Christmann, 2012).
  To be sure, the question of whether the sovereign people can never be subject to any limitation when it comes to decisions of passing constitutional amendments (Nugraha, 2022) cannot be answered with a resounding ‘yes’. In liberal democracies, the rule of law regulates and therefore sets some barriers to the exercise of popular sovereignty, especially in order to protect fundamental individual rights.
  The tensions between popular sovereignty and the rule of law are by no means a new phenomenon in direct democracy. In the case of Switzerland, for instance, the very first popular initiative that was adopted at the federal level clashed with fundamental individual rights. The ban on slaughtering, passed in 1893 in an anti-Semitic context, violated freedom of belief and conscience. While the people are basically sovereign only within the boundaries of the Constitution, new supranational restrictions have played an increasingly important role in recent decades. A significant development relates to the expansion of international law and, more generally, to a growing number of international commitments by nation states. To a certain degree, this has led to a detachment from domestic laws.
  In the European context, the embedding of human right protection within the European Convention on Human Rights (ECHR) has established such a supranational restriction for their member states. The noble assumption underlying this international protection is that human beings have an inherent dignity as well as equal and inalienable rights that no popular majority, no matter by what margin, should be allowed to overrule (Langer, 2010).
  The approval of ballot measures that clash with international law thus raises the question of the limits of popular sovereignty. A strict policy implementation would have a negative impact on international relations. Indeed, for the international legal system to work smoothly, it must rely on the primacy of international law over domestic law (see for instance Naef, 2017). This implies that ratified international treaties need to be observed at the national level. Compliance with international obligations is therefore based on the principle of pacta sunt servanda, which is customary in international law.
  When faced with a proposition that violates international law, an option for a national government would be of course to terminate or at least renegotiate the affected international treaties. Yet this is hardly feasible in many domains, especially when it comes to multilateral human right treaties such as the ECHR and the UN human rights pacts. Since states are under international obligation to respect human rights, it follows that they are not allowed to invoke referendum outcomes to justify an infringement (Nugraha, 2020).
  By borrowing on an idea developed by Mair (2013, 2014) in the domain of party politics, I argue that legislators (and other actors in charge of implementation) are typically confronted with a dilemma that revolves around responsiveness and responsibility. Responsiveness refers to the decisions-makers’ quality to express the policy preferences of citizens. Responsibility, for its part, is concerned with acting in line with legal and procedural rules and conventions that are anchored in the Constitution and international obligations of a given country.
  The dilemma that arises in the question at hand can be described as follows: if legislators act responsively by relying on a faithful implementation that contradicts international law, serious consequences may occur in the international realm. Condemnations by international courts may not only lead to reputational damages but also to the termination of treaties which are of crucial importance for a country. If, on the other hand, MPs show responsibility by implementing an initiative in conformity with international law, they expose themselves to the accusation of flouting the will of the people at the national level. In a nutshell, lawmakers face the obviously delicate task of squaring the circle between responsiveness and responsibility.
  In the context of Switzerland, which is at the centre of the empirical part of this article, this dilemma arises in connection with adopted popular initiatives that conflict with international law. At the federal level of the country, popular initiatives always refer to constitutional amendments. Given that the new constitutional provisions are usually not directly applicable, they require an implementation at the level of federal legislation. This task is the responsibility of the Federal Assembly (Parliament), which consists of the National Council (lower house) and the Council of States (upper house). In most cases, policy implementation of popular initiatives can thus be considered a political matter that falls into the hands of Parliament. The strong position of MPs is also reflected in the fact that there is no judicial review at the federal level. In addition, it follows from Article 190 of the Federal Constitution that the laws enacted by Parliament are valid even if they are unconstitutional. Taken together, this means that, at least theoretically, legislators have a considerable room of manoeuvre on implementation matters.
  Generally speaking, there is no remedy for resolving the conflict between obligations under international law and constitutional norms contrary to international law. To be sure, the implementation of such a popular initiative can be described as a highly political balance act that lacks specific legal guidelines in many respects (Naef, 2017). Article 5 Paragraph 4 of the Federal Constitution ensures a general primacy of international law over domestic law, without establishing clear rules in case of a conflict of norms, however. Furthermore, Article 190 of the Federal Constitution requires the Federal Supreme Court to apply both federal laws and international law, thereby opening the door for reconciling these two legislations.
  By contrast, the Federal Constitution is unambiguous regarding compliance with mandatory international law. This prohibition forms the only barrier for popular initiatives in relation to international law.²x Since there is no judicial review at the federal level of Switzerland, Parliament has the possibility to at least partially invalidate popular initiatives that violate mandatory provisions of international law before a nationwide vote is organised. This happens rarely though (Serdült, 2014, p. 73). In international comparison, the admissibility hurdles for popular initiatives are thus very low in the Swiss case (Marxer & Pállinger, 2009, p. 51). Yet the scope of mandatory international law is not clearly defined neither. The federal authorities have interpreted it to include ius cogens. These peremptory norms of international law cannot be derogated in any circumstances such as the prohibition of genocide, torture, slavery or inhuman and degrading treatment. They also include the principle of non-refoulement (i.e., the guarantee that no one should be (re-)turned to a country where they would face irreparable harm). By contrast, the remaining fundamental rights are not considered to be part of ius cogens. Accordingly, the vast majority of restrictions on fundamental rights do not constitute a limitation when it comes to the implementation of popular initiatives at the federal level of Switzerland.
  This crucial distinction between mandatory and non-mandatory international law implies that MPs will be reluctant to comply with provisions that are contrary to ius cogens. By contrast, there are no clear expectations as to demands of winning popular initiatives which, although contrary to international law, do not involve a violation of mandatory law. The policy outcome may vary as the case arises. In the realm of non-mandatory international law, it thus remains an open question whether the legislators of the Federal Assembly will make use of the possibility of watering down some problematic aspects of initiatives in order to respect Switzerland’s external obligations. Based on these considerations, I am now equipped to state the following hypothesis:
  H: Legislators will not implement provisions of accepted initiatives that conflict with mandatory international law (ius cogens), while they will be more open to comply with provisions that conflict with non-mandatory international law.

• 4 Case Selection

  The empirical analysis deals with Switzerland, a country that can be considered the paradigmatic case of direct democracy in the world. By contrast to referendums that prevent policy reforms from being enacted, popular initiatives allow for proposing new policies. While the success rate of the latter used to be very low (around 10%), this has slightly changed since the beginning of the twenty-first century. Among the 25 popular initiatives that have passed since 1893, 13 were accepted by a majority of both citizens and cantons (i.e., the Swiss member states) between March 2002 and June 2023. While their range includes a large number of issues, major implementation problems have occurred above all with those popular initiatives that conflict with international law (Hertig Randall, 2017). In recent years, many decision-makers and experts have indeed racked their brains as to how to comply with the popular mandate without disregarding Switzerland’s international obligations.
  It is no coincidence that this challenge takes place in current populist times. Indeed, popular initiatives that clash with international law are an integral part of the political strategy adopted by actors from the radical right, especially by the SVP, Switzerland’s biggest party (Bernhard et al., 2015; Biard, 2019). According to Hertig Randall (2017, 2018), the aim of this strategy consists in highlighting the encroachment of international norms on national sovereignty and to disregard both the international legal order and the rule of foreign judges, exemplified by the European Court of Human Rights (ECtHR).
  In order to gain a detailed understanding of the implementation process in the context of the federal level of Switzerland, I decided to conduct a single case study. Based on a stepwise procedure, I selected the popular initiative on the deportation of criminal foreigners (thereafter ‘deportation initiative’). Among the 13 popular initiatives that have passed thus far in the twenty-first century, seven cases can be considered to clash with some international law provisions. These popular initiatives are listed in Table 1 chronologically according to the date of acceptance.

  Table 1 Accepted popular initiatives at the federal level of Switzerland in the twenty-first century that conflict with international law (chronological order)

  Date	Proposition	Approval
  		Voters (%)	Cantons
  8 February 2004	Life imprisonment for dangerous offenders	56.2	21.5
  30 November 2008	Ban on statute of limitation for pornographic crimes	51.9	18
  29 November 2009	Minaret ban	57.5	19.5
  28 November 2010	Deportation of criminal foreigners	52.9	17.5
  9 February 2014	Against mass immigration	50.3	14.5
  18 May 2014	Prohibition of paedophiles to work with children	63.5	23
  7 March 2021	Burqa ban	51.2	18

  Note: For a popular initiative to pass a majority of both voters and cantons, n = 23 is required.

  As far as this pre-selection is concerned, the minaret ban as well as the burqa ban could not be taken into consideration here, since the former did not require a legislative bill to be enacted, while the implementation of the latter has not been completed yet. Subsequently, I decided to keep those popular initiatives that used to be disputed in terms of both mandatory and non-mandatory norms of international law. This is necessary in order to be able to examine whether, and if so, to what extent MPs of the Federal Assembly take into account this distinction in their implementation decisions (see Section 3). I thus eliminated the mass immigration initiative, given that only this case basically did not contain problematic provisions with respect to mandatory international law. Finally, I decided to set the initiatives’ authors as selection criteria by giving priority to parties over the remaining organisation. The reason for this choice relates to the fact that the preferences of parties are more likely to be accounted for than for usually less powerful and established collective actors, since MPs are among the implementation decision-makers. Among the four remaining cases, three popular initiatives originated from rather loosely organised citizen committees (i.e., the proposition on life imprisonment for dangerous offenders, the ban on statute of limitation for pornographic crimes and the prohibition of paedophiles to work with children). Hence, the remaining deportation initiative by the SVP is selected for this case study.

• 5 The Implementation of the Swiss Deportation Initiative

  The SVP launched the deportation initiative in the run-up to the 2007 federal elections. The proposition above all demanded criminal foreigners to be expelled from the country for a list of offences. The party managed to qualify this proposition to the ballot by collecting the required 100,000 signatures within 18 months with ease. Despite the fact that the deportation initiative clashed with various principles of international law (see below), Parliament decided not to invalidate it, considering that it may be possible to interpret its wording consistently with mandatory norms of international law in the case of success at the polls. On 28 November 2010, a majority of Swiss citizens (52.9%) and cantons (17.5 out of 23) came out in favour of the deportation initiative. On the same date, they defeated a so-called direct counter-proposal, by which both chambers of Parliament had tried to combat the deportation initiative. This alternative constitutional amendment was in conformity with international law, since it specifically complied with human rights and the proportionality principle by preventing automatic removal and enabling reviews in particular cases.
  As a result of the deportation initiative’s passage, Article 121 of the Federal Constitution was supplemented by four paragraphs. According to Paragraph 3, foreigners lose their right of residence and all other legal rights to remain in Switzerland if they are convicted of a range of offences (i.e., intentional homicide, rape or any other serious sexual offence, any other violent offence such as robbery, human trafficking or drug trafficking and burglary) or have improperly claimed social insurance or social assistance benefits. The convicted persons are to be subject to an entry ban of 5 to 15 years (Paragraph 5). According to transitional provisions, the Federal Assembly had to specify and complement the list of offences under Paragraph 3 and to enact penal provisions against persons who violate the entry ban within five years after the adoption of the new constitutional provisions.
  The deportation initiative clashes with international law in several ways, given that the aforementioned Paragraph 3 does not provide for any exception to automatic removal (Hertig Randall, 2018; Langer, 2010; Naef, 2017). Most importantly, there are some concerns with mandatory provisions of international law. According to Naef (2017, p. 257), there is a conflict with Article 3 of the ECHR, which prohibits torture, inhuman or degrading treatment and punishment. In recent practice, the ECtHR (and the United Nations Human Rights Committee) also established a violation of the right to life in Article 2 of the ECHR when a person facing a death sentence is expelled to that country. In the same way, the non-refoulement principle that is enshrined in the Geneva Convention relating to the Status of Refugees is affected.
  Regarding non-mandatory provisions of international law, conflicts arise with the right to respect for private and family life in Article 8 of the ECHR, which requires a case-by-case proportionality analysis (Hertig Randall, 2018, p. 771) as well as with the Convention on the Rights of the Child. In addition, there is an inconsistency with the Agreement on the Free Movement of Persons (AFMP) between Switzerland and the European Union. The automatic deportation mechanism violates a directive of the EU according to which the AFMP can only be restricted for reasons of public order, safety and health. A faithful implementation could not only jeopardise this specific treaty as a result of a conceivable unilateral termination by the EU but, due to the so-called guillotine clause, six sectoral bilateral agreements in the domains of agriculture, research, public procurement, technical barriers to trade, land transportation and civil aviation as well.
  In the following, I shall track the implementation process of the deportation initiative. To that end, this case study relies on a systematic reading of the Neue Zürcher Zeitung, a Swiss high-quality newspaper. Based on the search term ‘deportation*’ (Ausschaffung* in German) those 36 articles that primarily addressed the implementation of the popular initiative under scrutiny from November 2010 to December 2015 were selected as a source for this case study.
  The implementation process of the deportation initiative was in line with common Swiss practice for those cases there is a need for legislation at the federal level. The remainder of this section is thus divided into two parts. The first part is devoted to the pre-parliamentary phase (working group, consultation procedure and governmental draft submitted to both chambers of Parliament), which is of importance as it pre-structures the bargaining in the Federal Assembly. The second part deals with the implementation decisions made by both chambers of Parliament. Given that they have equal rights, the National Council and the Council of States had to reach an agreement in order to enact legislative bills. Each chamber can thus be regarded as a veto player. It is important to note that the MPs of the SVP were far from reaching majority in both houses.³x During the 49th legislative term of the Federal Assembly (2011-2015), in which the deportation initiative was adopted, the SVP had 59 out of the 200 seats of the National Council and only 7 out of the 46 seats of the Council of States. Hence, the supporters of the deportation initiative were dependent on other parties in order to achieve a strict implementation.

  5.1 Pre-Parliamentary Phase

  While the Federal Assembly was in charge of implementing the new constitutional provisions, the government (i.e., the Federal Council) had first to decide on a legislative proposal it would submit to the two chambers of Parliament. To that end, Justice Minister Simonetta Sommaruga set up a working group in December 2010 with the mandate to elaborate some implementation proposals. The seven-member panel was headed by Heinrich Koller, a law professor and a former director of the Federal Office of Justice. In addition to respectively two representatives of the federal administration and the cantons, it included two members of the Swiss People’s Party (SVP).
  Unsurprisingly, the working group did not manage to agree on a joint proposal. A majority of experts recommended implementing the initiative in a modified way, considering the automatism demanded by the initiative to leave no room for judging individual cases and would therefore lead to Switzerland being condemned by the European Court of Justice. In addition, the majority feared a revocation of AFMP by the European Union and thus of six additional sectoral bilateral treaties. By contrast, the representatives of the SVP insisted on sticking to the initiative’s wording.
  In its final report of June 2011, the working group presented four implementation variants. In all variants, the working group committed itself to compliance with the principle of non-refoulement, which is part of ius cogens. The first variant, favoured by the SVP and rejected by the majority of the working group, included an automatic deportation of foreigners after a conviction for an offence mentioned in the constitutional provision. In addition to social abuse and burglary, the catalogue of offences that would result in deportation focused on everyday crimes of violence and sexual offences. In the absence of a minimum penalty as a prerequisite for deportation, a significantly higher number of offenders would have been affected than under the three remaining variants. These provided for various exceptions depending on the severity of penalty and the nature of offence.
  The majority of experts pleaded for the second variant, under which the list of offences included the crimes that are explicitly mentioned in the new constitutional paragraph as well as other serious crimes. For these offences, penalties of at least six months would lead to deportation. The second variant basically corresponded to the counter-proposal citizens had rejected on 28 November 2010. The only notable exception was a tightening of the minimum penalty length, which was reduced by six months.
  A majority of the working group proposed the third and fourth variants as alternatives to the second one. Under the third variant, the catalogue of offences contained crimes against specific legal interest such as ‘life and limb’, ‘freedom’, ‘sexual integrity’ as well as crimes against public peace. The fourth variant, for its part, proposed an implementation in the framework of the law on foreigners (as opposed to the penal legislation) by a mandatory revocation of permits and a mandatory entry ban based on the requirements of the second variant.
  The SVP was quick to react. By proposing a minimum penalty length for deporting criminal foreigners, party president Toni Brunner declared that the majority of the working group had blatantly disregarded the will of the people. But more drama was to come. Given that the party anticipated that the deportation initiative would be watered down by Parliament, it decided to launch another popular initiative that aimed at enshrining the party’s desired legislative amendment in the Constitution. The so-called enforcement initiative was launched in July 2012, without waiting for the Federal Council to submit its proposal to the two chambers of Parliament. With this strategic move, the SVP managed to put some pressure on the other parties so that they would finally accept a strict implementation of the deportation initiative. This enforcement initiative listed numerous offences that would lead to deportation, including inflicting bodily injury, kidnapping, public incitement to crime or violence, counterfeiting currency or committing sexual offences against children. In addition, it contained a definition of peremptory norms of international law by limiting them down to the prohibition of torture, genocide, aggression, slavery as well as to the principle of non-refoulement.⁴x The Federal Assembly decided to invalidate this particular provision in March 2015.
  Based on the proposals made by the working group, the Federal Council put forward two drafts for consideration by relevant stakeholders (such as parties, interest groups and state actors) in May 2012. In the framework of a consultation procedure about legislative amendments of the penal code and the military penal code, the Federal Council made it clear that it favoured the first draft. Based on the new constitutional provision, this variant set out a list of specific offences and expanded the catalogue to include serious crimes against property. In principle, a convicted foreign criminal would be expelled from Switzerland if he or she is sentenced to at least six months of imprisonment. Government proposed an exception for those deportations that would lead to a serious violation of international human rights guarantees. Whereas the minimum penalty would make sure that automatic deportation does not apply to minor offences, criminal tourists and repeated offenders would nevertheless be expelled for a minimum of five years for minor offences.
  The second draft stemmed from the SVP’s representatives of the working group. It not only contained a broader list of offences (especially in the area of violent offences to include lighter crimes and misdemeanours such as simple bodily harm) but also provided for deportations even for petty offences, given that judges would always have to pronounce a mandatory expulsion, regardless of the sentence pronounced.
  By pleading for not applying the principle of automatic deportation for prison sentences of less than six months, the Federal Council sought to take into account the constitutional principle of proportionality and to minimise incompatibilities with international law. According to its assessment, the first draft would largely comply with the ECHR and the UN Pact II, even though it was aware that the first draft would not fully meet these requirements. Deportations based on the provisions proposed by the Federal Council could also violate the AFMP with the European Union. Indeed, the relevant EU law under the Agreement requires a case-by-case examination, which would not always be compatible with the new constitutional provision. However, there was no doubt that the first draft would better account for these provisions than the second draft, since the former restricted the list of offences to serious crimes, provided for a minimum penalty and included the human rights guarantees under international law on deportation matters.
  The vast majority of participants in the consultation procedure pronounced themselves in favour of the first draft. It was therefore no surprise that the draft of the Federal Council submitted to Parliament in June 2013, was strongly oriented towards this variant. However, compared with the initial draft, the catalogue of offences was expanded to include, among other things, offences in the area of social welfare fraud. In its report, the Federal Council explicitly mentioned that international human rights guarantees had to be respected. It thus proposed to include a provision in the penal code (Article 66d) that explicitly requires mandatory international law to be taken into account when it comes to enforcing deportations. As for non-mandatory provisions, the government drew attention to some dangers of not conforming to international law. It stated that recent constitutional law may take precedence over conflicting international law. As a result, the proposed implementation would lead to conflicts with the ECHR, the UN Pact II and the AFMP.
  These rather concerning considerations about a possible non-compliance with international law provisions had to be appreciated in the context of a controversial decision taken by the Federal Supreme Court in October 2012 and published in February 2013. As expected, the court denied direct applicability of the new constitutional article, thereby arguing that it was the duty of the Federal Assembly to define the relationship of the deportation initiative with international law. Yet in an obiter dictum, the Federal Supreme Court additionally held that in case of an irreducible conflict, precedence was to be given to the ECHR over the constitutional provision. This rather unequivocal statement caused quite a stir, since it put some pressure on Parliament to comply with international law as far as the implementation of the initiative at hand was concerned.⁵x Beyond that, the obiter dictum produced a considerable political backlash, leading to the launching of yet a popular initiative by the SVP. The so-called “self-determination initiative” sought to establish a precedent of the Federal Constitution over any other law except for peremptory norms of international law (ius cogens).

  5.2 Parliamentary Phase

  Among the two chambers of Parliament, the National Council was the first to consider the implementation of the deportation initiative in March 2014. The political institutions committee of the lower chamber had previously asked the federal administration to elaborate a variant that was based on the SVP’s enforcement initiative. It turned out that the representatives of the major parties from the right (i.e., the Swiss People’s Party, the Liberals, (FDP) and the Christian Democrats, (CVP)) favoured this approach over the governmental draft. Philipp Müller, the president of the FDP was the driving force behind the adoption of this harder line. Müller’s goal was to take the wind out of the sails of the SVP. He was determined to prevent another ballot on criminal foreigners in the context of the 2015 federal elections. In line with many MPs from the moderate right, Müller was convinced that people and cantons would also come out in favour of the enforcement initiative in the framework of a nation-wide vote and that it was therefore preferable to have some controversial provisions stipulated in the penal law than in the Federal Constitution.
  In the plenary, a majority of 106 votes against 65 accepted a bill that made substantive concessions to the SVP. Most importantly, the National Council decided that a minimum penalty would not be a prerequisite for deportation regarding an extended list of offences. Such an automatic mechanism would apply to intentional homicide, murder, manslaughter, grievous bodily harm, qualified theft, robbery, commercial fraud, sexual assault, rape, drug offences and social welfare fraud. For other offences, such as simple assault, deportation would only occur in the case of repeated offences.
  Opposition came only from the Social Democrats and the Greens as well as from a majority of the Green Liberals. These MPs pointed out in vain that back in 1999, people had approved the Federal Constitution that explicitly contains the principle of proportionality. In this context, Ms. Sommaruga observed that the fundamental guarantees of the rule of law had not been abolished with the adoption of the deportation initiative. Several National Councillors from the right rejected the governmental approach to implementation on the grounds that it contained central provisions of the counter-proposal citizens had rejected in November 2010.
  In light of this strict implementation, Mr. Brunner announced that the SVP would possibly withdraw the enforcement initiative in the case the Council of States would follow the lower chamber in the most important points. However, due to major constitutional and international law concerns, there was a big question mark behind the fulfilment of this condition indeed. While complying with mandatory international law in the enforcement of deportations by adopting Article 66d of the penal code, the National Council’s implementation of the deportation initiative severely undermined the constitutionally guaranteed right to proportionality and some non-mandatory provisions of international law. In individual cases, it would have been therefore left to the courts to weigh the matter accordingly. It was foreseeable that there would also be conflicts with the ECHR, with the ECtHR in Strasbourg taking corrective action.
  In June 2014, the political institutions committee of the Council of States announced that it wanted to take its own path of implementation by searching for a middle ground between the Federal Council’s and the National Council’s proposals. Yet the committee struggled with elaborating an appropriate solution that was supported by a larger majority of its members. It was only in November that it presented a new draft. Contrary to the lower chamber, it did not base its proposal on the enforcement initiative but on the deportation initiative. Accordingly, all offences listed there provided for mandatory deportation. This would also apply to welfare abuses as well as to some serious crimes such as forced marriage and genital mutilation that had not been taken into account thus far. For other offences, however, a more differentiated approach was envisaged. More specifically, mandatory deportation should not apply to misdemeanours and offences of application.
  Yet the core of the proposal referred to a hardship clause. Under this provision, courts would have the possibility to waive deportations in case of ‘serious personal hardship’ and if the public interest does not outweigh the private interests of the person concerned. Special consideration would have to be given to the situation of foreigners who were born and raised in Switzerland (i.e., so-called Secondos).
  In December 2014, this bill passed with ease in the upper house. The Senators followed the recommendation of their committee by 28 votes to 3. Although some of their moderate members had accepted the draft in the Council of States, the SVP was not satisfied with the outcome. The party was particularly bothered by the hardship clause, considering that the people had spoken out in favour of automatic deportations. The supporters argued that such a clause was necessary in order to maintain the principle of proportionality. They also emphasised that this clause would only come into play under very limited conditions.
  After the decision taken by the Council of States, observers of Swiss politics agreed that this compromise solution would be likely to be adopted by a majority of the National Council. This was primarily due to members of the CVP and the FDP who were expected to change sides. In March 2015, two proposals by the SVP to stick to the National Council’s initial implementation draft and to at least not include the hardship clause failed by clear margins. With only some minor changes, the MPs of the lower chamber agreed on the Council of States’ version. The Senators subsequently approved the legislative they obtained from the National Council, thereby paving the way for the final votes.
  The Council of States adopted the legislative bill by 36 votes to 3 with 5 abstentions. The nays came all from the SVP’s parliamentary group. The National Council, for its part, voted in favour by 109 votes to 68 with 18 abstentions. In addition to the united radical right (i.e., SVP, Ticino League and Geneva Citizens’ Movement), three MPs of the FDP as well as the majority of the Greens rejected the bill. The rather surprising opposition from the left was attributable to the fact that the Greens were fundamentally opposed to the deportation of foreigners from the second and the third generation. The implementation of the deportation initiative eventually came into force in October 2016 seven months after which people and cantons had rejected the enforcement initiative at the ballot box, not least due to a strong mobilisation by civil society.
  To sum up, the implementation of the deportation initiative turned out to be a stony and controversial journey onto Swiss penal legislation. After a long process that included a lot of stakeholders, the Federal Assembly enacted a decisive tightening in deportation matters that essentially proved to be in line with the initiative’s aim. Most importantly, it specified and extended the list of offences that leads to deportations. At the same time, Parliament basically managed to comply with Switzerland’s international obligations, above all with respect to ius cogens. This was achieved by anchoring the consideration of mandatory international law as far as enforcement decisions are concerned in a new article of the penal code. Regarding concerns about both mandatory and non-mandatory provisions of international law, the introduction of the hardship clause proved to be of crucial importance. Under the decisive influence of the Council of States, commonly referred to as la chambre de réflexion (i.e., the chamber of reflection) in Swiss politics, this clause enables judges to justify exceptions to automatic deportations. However, it needs to be mentioned that this approach falls short of a full proportionality review (Hertig Randall 2017, p. 135), thus not conforming to a EU directive on the AFMP for instance.⁶x These incompatibilities between federal legislation and supranational law may be the subject of court rulings relating to individual cases of deportation decisions. Hence, compliance with some specific non-mandatory provisions of international law was not completely accomplished. Overall, the results are in line with the hypothesis, given that it turned out that MPs refrained from implementing provisions that clashed with ius cogens, while being less strict regarding when it came to non-mandatory international law.

• 6 Conclusion

  Initiatives play an important role in direct democracy, given that they enable weak actors to submit new policies to citizens. Once such propositions pass at the ballot box, implementation is still pending, however. When it comes to transforming winning initiatives into policy change, the conclusion reached by scholars thus far proves to be rather sobering. Indeed, many accepted initiatives have been found not to take full effect or even no effect at all. To explain this intriguing finding, political scientists have focused on leadership. From this perspective, widespread non-compliance arises from the fact that actors in charge of implementation are typically reluctant to meet the demands of initiatives approved by the people because they basically dislike them.
  Yet the political science literature on policy implementation in direct democracy has somewhat neglected the fact that in liberal democracies, non-compliance by decision-makers can be the result of supranational constraints. In an attempt to advance the existing literature, this article has focused on the role played by international law in legislators’ responses to the passage of initiatives. I have argued that the actors in charge of implementation face a fundamental dilemma between responsiveness and responsibility in such situations. The former refers to following a people’s will by complying with the demands of an initiative, while the latter consists in not adopting a faithful implementation in order to respect a country’s external legal obligations. Against the background of the Swiss context, this article has focused on the distinction between mandatory and non-mandatory international law. I have hypothesised that legislators do not implement provisions that conflict with the former, while they are more open to comply with provisions that conflict with the latter.
  The empirical part of this article has dealt with a case study of the Swiss deportation initiative, a proposition that conflicted with both mandatory and non-mandatory international law. The summary of this long and rather complex implementation process that included a multitude of stakeholders has shown that the Federal Assembly was eventually close to squaring the circle. Indeed, the amendments of the penal code basically met the demands of the popular initiative and at the same time respected most concerns regarding international law. More specifically, the latter was above all achieved through the introduction of a hardship clause that provides judges with the possibility to justify exceptions to automatic deportations. In line with the hypothesis tested here, the mandatory norms of international law (ius cogens) were fully complied with, while this proved not to be completely the case as far as some specific issues relating to non-mandatory international law are concerned.
  It can be stated that Parliament eventually did a fairly good job in balancing out the initiative’s demands with international law obligations, given that the translation of the new constitutional provisions into federal law was watered down only to the extent that was necessary. Despite the fact that the initiators from the Swiss People’s Party expressed dissatisfaction with the implementation outcome, it appears not very convincing to argue that legislators had disregarded the people’s will. In this context, it needs to be highlighted that the SVP achieved a lot with the deportation initiative in all stages of policy-making, ultimately leading the Federal Assembly to enact a decisive tightening in deportation matters (Bernhard et al., 2021; Biard, 2017). Altogether, MPs thus displayed a great deal of responsiveness, while showing responsibility with respect to international law. This case study illustrates that though power emanates from the people in a democracy, popular sovereignty cannot be absolute, especially when compliance with fundamental principles of superior law is at stake.
  However, it remains an open question whether Parliament would have adopted such a faithful implementation in other political circumstances. The fact that the SVP put a lot of pressure on MPs, especially by launching an enforcement initiative during the implementation process certainly contributed to a high degree of compliance in the case of the deportation initiative. The Federal Supreme Court also played a significant role, as it urged legislators to act responsibly by complying with international law. It needs to be mentioned that this rather surprising intervention is of importance well beyond the case at hand. Hence, scholars may thus want to focus on the implementation of other initiatives that clash with international law. To that end, the context of Switzerland provides researchers with numerous opportunities. In this context, a thus far neglected aspect in the processes of implementation is international and supranational actors. Indeed, EU or UN organisations can be expected to attempt to influence the Swiss decision-makers to favour responsible policy outcomes that are in conformity with international law, be it in the intergovernmental sphere or via lobbying.
  Yet scholars may also put into perspective the Swiss evidence to the experiences made in the numerous countries where ballot measures have played an increasingly important role in recent years. While the literature on policy implementation in direct democracy is generally underdeveloped, the lack of internationally comparative work can be considered a major blind spot indeed.

References

• Bali, V. A. (2003). Implementing Popular Initiatives: What Matters for Compliance? The Journal of Politics, 65(4), 1130-1146.

• Bernhard, L., Biancalana, C. & Mazzoleni, O. (2021). Switzerland. In D. Albertazzi & D. Vampa (Eds.), Populism and New Patterns of Political Competition in Western Europe (pp. 148-167). London: Routledge.

• Bernhard, L., Kriesi, H. & Weber, E. (2015). The Populist Discourse of the Swiss People’s Party. In H. Kriesi & T. S. Pappas (Eds.), European Populism in the Shadow of the Great Recession (pp. 123-137). Colchester: ECPR Press.

• Biard, B. (2017). The Influence of the Swiss SVP on Policy-Making: Opening the Black Box. Czech Journal of Political Science, 24(1), 21-36. doi: 10.5817/PC2017-1-21.

• Biard, B. (2019). The Influence of Radical Right Populist Parties on Law and Order Policy-Making. Policy Studies, 40(1), 40-57. doi: 10.1080/01442872.2018.1533110.

• Christmann, A. (2012). Direct Democracy and the Rule of Law: Assessing a Tense Relationship. In W. Marxer (Ed.), Direct Democracy and Minorities (pp. 47-63). Springer VS.

• Ferraiolo, K. (2007). Preserving the Initiative: State Legislative Response to Direct Democracy. Polity, 39(4), 425-448. doi: 10.1057/palgrave.polity.2300087.

• Gerber, E. R. (1999). The Populist Paradox: Interest Group Influence and the Promise of Direct Legislation. Princeton University Press.

• Gerber, E. R., Lupia, A. & McCubbins, M. D. (2004). When Does Government Limit the Impact of Voter Initiatives? The Politics of Implementation and Enforcement. The Journal of Politics, 66(1), 43-68. doi: 10.1046/j.1468-2508.2004.00141.x.

• Gerber, E. R., Lupia, A., McCubbins, M. D. & Kiewiet, D. R. (2001). Stealing the Initiative: How State Government Responds to Direct Democracy. Prentice Hall.

• Hertig Randall, M. (2017). Direct Democracy in Switzerland: Trends, Challenges and the Quest for Solutions. In A. Chommeloux & E. Gibson-Morgan (Eds.), Voting in Contemporary Europe (pp. 129-156). Palgrave Macmillan.

• Hertig Randall, M. (2018). Taming the Demos Through Courts? The Example of the Swiss Deportation Initiative. Quaderni Costituzionali, 38(3), 765-776.

• Jäske, M. & Setälä, M. (2019). Referendums and Citizens’ Initiatives. In S. Elstub and O. Escobar (Eds.), Handbook of Democratic Innovations and Governance (pp. 90-104). Edward Elgar.

• Langer, L. (2010). Panacea or Pathetic Fallacy? The Swiss Ban on Minarets. Journal of Transnational Law, 43(4), 863-951.

• Lupia, A. & Matsusaka, J. G. (2004). Direct Democracy: New Approaches to Old Questions. Annual Review of Political Science, 7, 463-482. doi: 10.1146/annurev.polisci.7.012003.104730.

• Mair, P. (2013). Smaghi Versus The Parties: Representative Government and Institutional Constraints. In A. Schäfer & W. Streeck (Eds.), Politics in the Age of Austerity (pp. 143-168). Polity Press.

• Mair, P. (2014). Representative Versus Responsible Government. In P. Mair (Ed.), On Parties, Party Systems and Democracy (pp. 581-596). ECPR Press.

• Marxer, W. & Pállinger, Z. T. (2007). System Context and System Effects of Direct Democracy: Direct Democracy in Liechtenstein and Switzerland Compared. In Z. T. Pállinger, B. Kaufmann, W. Marxer & T. Schiller (Eds.), Direct Democracy in Europe: Developments and Prospects (pp. 12-29). VS Verlag für Sozialwissenschaften.

• Marxer, W. & Pállinger, Z. T. (2009). Stabilizing or Destabilizing? Direct-Democratic Instruments in Different Political Systems. In M. Setälä & T. Schiller (Eds.), Referendums and Representative Democracy: Responsiveness, Accountability and Deliberation (pp. 34-55). Routledge.

• Miller, K. P. (2009). Direct Democracy and the Courts. Cambridge University Press.

• Naef, T. (2017). Popular Initiatives Contrary to International Law: A Swiss Dilemma. In A. Kellerhals & T. Baumgartner (Eds.), Rule of Law in Europe (pp. 241-271). Schulthess.

• Nugraha, I. Y. (2020). Human Rights Referendum: Dissonance Between “The Will of the People” and Fundamental Rights? Netherlands Quarterly of Human Rights, 38(2), 115-133.

• Nugraha, I. Y. (2022). Popular Sovereignty and Constitutional Referendum: Can “The People” Be Limited by Human Rights? German Law Journal, 23(1), 19-43. doi: 10.1017/glj.2022.2.

• Schacter, J. S. (1995). The Pursuit of “Popular Intent”: Interpretive Dilemmas in Direct Democracy. The Yale Law Journal, 105(1), 107-176.

• Serdült, U. (2014). Referendums in Switzerland. In M. Qvortrup (Ed.), Referendums Around the World: The Continued Growth of Direct Democracy (pp. 65-93). Palgrave Macmillan.

• Setälä, M. (1999). Referendums in Western Europe: A Wave of Direct Democracy? Scandinavian Political Studies, 22(4), 327-340.

Noten

• 1 A reviewer of this article convincingly pointed to an additional crucial institutional characteristic from a theoretical perspective by drawing the distinction between legally binding and advisory direct-democratic votes (Setälä, 1999, p. 338). The former can be reasonably expected to be more strictly implemented than the latter. However, advisory votes may in fact also provide strong guidance to decision-makers (Jäske & Setälä, 2019, p. 91).

• 2 Since there is no judicial review at the federal level of Switzerland, Parliament has the possibility to at least partially invalidate popular initiatives that violate mandatory provisions of international law before a nationwide vote is organised. This happens rarely though (Serdült, 2014, p. 73). In international comparison, the admissibility hurdles for popular initiatives are thus very low in the Swiss case (Marxer & Pállinger, 2009, p. 51).

• 3 During the 49th legislative term of the Federal Assembly (2011-2015), in which the deportation initiative was adopted, the SVP had 59 out of the 200 seats of the National Council and only 7 out of the 46 seats of the Council of States.

• 4 The Federal Assembly decided to invalidate this particular provision in March 2015.

• 5 Beyond that, the obiter dictum produced a considerable political backlash, leading to the launching of yet a popular initiative by the SVP. The so-called “self-determination initiative” sought to establish a precedent of the Federal Constitution over any other law except for peremptory norms of international law (ius cogens).

• 6 These incompatibilities between federal legislation and supranational law may be the subject of court rulings relating to individual cases of deportation decisions.