Draft: Please do not cite or circulate without permission. Criticisms and comments most welcome! Abstract

Transcription

1 Political representation in semi-direct democracy: The impact of the facultative referendum Alice el-wakil, PhD Candidate, University of Zurich / Centre for Democracy Studies Aarau International Political Science Association World Congress Poznan, July 23-28, 2016 Panel RC34.04, Democracy: Representation, Participation, Contestation Draft: Please do not cite or circulate without permission. Criticisms and comments most welcome! Abstract Most normative and empirical defenses of direct democratic procedures emphasize the way in which they supplement the representative institutions. They are framed as some sort of add-ons to representative institutions, which are themselves taken as an unchanged, and unchangeable background. This approach might have caused research to overlook the potential changes that direct institutions bring to representative ones; as Simon Hug notes, little research has addressed the issue of how referendums affect the representative governments that prevail in all democratic states around the world (Hug 2008, 252). In this paper, I propose to contribute to filling this gap by looking at what the presence of direct democratic institutions in a representative system changes, at the theoretical level, in the latter rather than at what it brings to it. I do this by clarifying the impact of one specific bottom-up direct institution, the facultative referendum, on the conception of representation and of the role of representatives. I first present the additional authorization and accountability mechanisms this institution introduces, and how it enables the emergence of democratic unelected representatives. I then argue that its presence makes it impossible for representatives to act as trustees. This doesn t mean that trust is eliminated from the semi-direct political system altogether: The presence of the facultative referendum rather seems to be able to promote a view of selective trust in representative institution on an issue-by-issue basis, which encourages citizens to actively engage in judging their representatives work. Keywords: Contestation, popular vote, selective trust, unelected representation 1. Introduction While contemporary theorists of democracy have mostly favored representative political systems over more direct ones to implement their various democratic ideals, some researchers have recommended the introduction of certain bottom-up direct democratic procedures within representative systems in order to enhance the quality of existing democracies (see e.g., Setälä 2006; Steenbergen 2009; Parkinson 2009). Direct democratic procedures could be considered to play a valuable role in terms of encouraging public deliberation on specific issues (Setälä 2006, 703), of countering the alienation emanating from more conventional electoral processes that citizens might experience (Gastil, Richards, and Knobloch 2014, 80), of bringing the preferences of the representatives closer to the ones of the citizens (Leemann and Wasserfallen tbp; Sager and Bühlmann 2009, 201), of raising the level of trust in authorities (Bauer and Fatke 2014), or of increasing the legitimacy of adopted policies (Hug 2008, 256;

2 Chambers 1998) once we take on board the need for electoral chambers of representatives, and an array of administrative and enabling bodies for it to oversee (Saward 2001, 371; see also Qvortrup 2015, 44). These defenses of direct democratic procedures emphasize the way in which they supplement the representative institutions. They are framed as some sort of add-ons to representative institutions, which are themselves taken as an unchanged, and unchangeable background (Setälä and Schiller 2009, 6). 1 This approach might have caused research to overlook the potential changes that direct institutions bring to representative ones; as Simon Hug notes, little research has addressed the issue of how referendums affect the representative governments that prevail in all democratic states around the world (Hug 2008, 252). 2 In this paper, I want to contribute to filling this gap by changing outlook, namely by looking at what the introduction of direct democratic institutions in a representative system changes, at the theoretical level, in the latter rather than at what it brings to it. More precisely, I aim at clarifying the impact of one specific bottom-up direct institution, the facultative referendum, on the conception of representation and of the role of representatives. Clarifying these changes in the representative, background institutions and evaluating them is an essential step towards offering a convincing defense of semi-direct (or mixed) systems. As such, it could inform the fundamental choice [ ] between a pure representative system and a similar system combined with referendums (Hug 2008, 252; see also Cheneval 2015, 169) as well as open up the possibility for new reflections about representation. Before starting with the argument, let me clarify what is the facultative referendum, and why I focus on this specific institution. The facultative referendum is a bottom-up direct democratic institution, which is implemented in Switzerland at the national level, that enables citizens to contest almost any law adopted by their representatives. In its first, referendum launching phase, they have to gather the required number of signatures in the specified amount of time. In a second step, a popular vote is organized and the majority of citizens decides, in a binding way, on the adoption or the rejection of that law. Focusing on this direct democratic mechanism only benefits the present paper in three ways. First, the facultative referendum institution corresponds to the kind of add-ons recommended by most proponents 1 The question guiding their research is often in what way can referendums supplement or undermine representative democracy? a question to which many answer that the direct democratic institutions could play the role of democratic correctives (Leemann and Wasserfallen tbp; see also Sager and Bühlmann 2009, 201) to be used as supplements to the existing institutions of representative democracy (Setälä 2006, ). 2 In his view, proponents of referendums fail to address convincingly the interactions referendums create between the representative system and the elements of direct citizen participation (Hug 2008, 252). 2

3 of direct democratic institutions. 3 Second, limiting my argument to one specific direct democratic institution prevents me from the risk of blurring the relevant differences between different direct democratic political institutions, 4 and represents a way to gain in argumentative precision and clarity. 5 Third, this analysis can open the way to a more general claim about the impact of other direct democratic mechanisms on representative institutions. If even the facultative referendum, an institution that presupposes a representative institutional context, modifies the conception of representation in relevant ways, other more controversial direct democratic institutions (such as the popular initiative) might also have an impact on these background institutions that would deserve to be made explicit and evaluated. The paper is divided into three parts. In the first one, I propose a translation of the facultative referendum procedure in the terms used by theories of representation. I offer a table that highlights the different mechanisms of authorization and accountability, but also the different kinds of democratic representatives present in a representative democracy with facultative referendum procedures. In the second part, I argue that the presence of this institution makes it impossible for representatives to act as trustees. However, I try to show in the third part that, instead of transforming representatives in mere delegates, the facultative referendum seems to contain a notion of trust that corresponds to a more recent view of trust in representatives a view that answers the radical worry of disengagement of the citizens in the process of politics, which undermines the accountability of democratic representatives (Cohen and Fung 2004; Barber 1984). By providing an institutional tool for citizens to actually check the decisions made by representatives, and giving them an incentive to continuously judge their actions and make the decision to trust their representatives or not (Warren and Gastil 2015, 564), the facultative referendum right encourages elected representatives to be prepared to justify their decisions to a wide audience and to act in an responsive and anticipatory way. I conclude by summarizing what this reversed approach tells us about the impact of the facultative referendum on the representative institutions and by highlighting the valuable character of the semi-direct political system created. 3 Maija Setälä, who calls this process rejective referendum (Setälä and Schiller 2009, 5) or decisioncontrolling referendum (Setälä 2006, 714), for instance strongly supports it (Setälä and Schiller 2009, ). 4 For instance, while John Gastil and Robert Richards claim to highlight the weaknesses of direct democratic mechanisms, i.e. referenda and initiatives, two of their main criticisms to it are actually only directed to the popular initiative (Gastil and Richards 2013, 255); in the same way, their proposal for introducing random assemblies in the direct democratic process appears to only focus on initiative procedures (Gastil and Richards 2013, 266). 5 Even more so after the recent and almost unanimously criticised BREXIT referendum in Great Britain, which differs from the facultative referendum procedure in important and normatively relevant ways. 3

4 1. The facultative referendum in representative terms Direct democratic institutions are rarely accounted for in the same terminology as representative ones. However, this vocabulary appears to be to a certain extent appropriate to highlight unexplored aspects of the facultative referendum is about, and better understand how it impacts the elected representatives. In all representative democracies, including the ones with provisions for direct mechanisms, representatives are selected by the citizens in order to make binding political decisions for the society on the entire bundle of issues upon which they can legislate (Besley and Coate 2008). The democratic theory literature mentions two conditions for representatives to be democratic ones (Pitkin 1967; Urbinati and Warren 2008, 396). On the one hand, they must be authorized by the citizens to take action in a certain domain. This authorization moment is a process of gaining power for representatives-to-be. On the other hand, citizens must have ways to hold them accountable for their actions in this domain. This second feature is said to guarantee some level of responsiveness on the part of the representative (Urbinati and Warren 2008, 402). Fair elections are the main institution recommended to fulfill these two criteria of democratic representation. 6 It is through elections that citizens, by casting their votes, select representatives and authorize them to make decisions for their constituency for a determined period of time. And it is through elections that they can hold representatives accountable for their actions during their entire term and, eventually, sanction them by not re-electing them (Hug 2008, 260). In this sense, regular elections are the main check to avoid gross abuses of power (Urbinati and Warren 2008, 402). The introduction of the facultative referendum in a representative democracy adds an authorization and an accountability mechanism, which can be used during the electoral term. By collecting the required number of signature within the allocated time, a dissatisfied minority 7 can temporarily de-authorize the representatives to make decisions for the constituency on the specific law subjected to the referendum. When a facultative referendum is launched, representatives loose their competence to represent the constituency on that issue. This withdrawal of authorization lasts until the popular vote takes place; right after it, and 6 That is, fair elections in a democratic context. As Nadia Urbinati insists, elections take place in a context with robust local autonomy and freedom of speech and association as well as some basic equality of material conditions to enable democratic representation (Urbinati 2006, 39). 7 In Switzerland, 50'000 signatures must be collected within 100 days to launch a referendum. 4

5 independently of whether the majority of citizens has supported or rejected the law, the representatives are automatically authorized to make decisions on this issue again. The popular vote thus constitutes the additional accountability mechanism introduced by the facultative referendum. At this stage, citizens choose to support either the representatives by accepting the law proposal, or the referendum committee, by favoring the status quo. In the latter case, the majority of citizens sanction the representatives in a soft way, as Alexandre Trechsel argues, since the accountor keeps the quality of accountor, i.e. remains in office (or keeps the right to vote) but their politics are potentially affected by the sanction and they might change a political decision or express a different political opinion (Trechsel 2010, 1059). These authorization and accountability mechanisms only concern a single issue, one among the many law projects adopted by representatives. Representatives are de-authorized and held accountable only for this specific issue; they keep their competence to make democratic decisions about all the other ones. In this sense, the facultative referendum makes it possible to unbundle the issues upon which representatives can take action (Besley and Coate 2008, 2). 8 While the election only makes it possible to punish representatives for their overall action on all issues during an entire term, the facultative referendum enables citizens to hold their representative accountable on single issues (Trechsel 2010, 1055). 9 The facultative referendum introduces an additional specificity: It enables the emergence of representatives outside electoral cycles (Saward 2009, 4). The launchers of a specific facultative referendum procedure actually become some sort of unelected representatives. According to institutional definitions of representation, the referendum launchers are representatives in the sense that they are able to influence the political decision-making procedure directly and for the entire society by asking that the last word on a specific law be given to the citizens instead of the representatives. The facultative referendum procedure also provides them with visibility to argue and campaign for this law s rejection before the popular vote. According to broader definitions of representation, such as the one Michael Saward offers, the referendum launchers are representatives to the extent that they make claims to be representatives (Saward 2006, 298). By launching a referendum, simple citizens appear to make two kinds of claims: A general one, about minorities or majorities whose interests are not served by the contested law, and a more specific one, deriving from the signature 8 Timothy Besley and Stephen Coate direct this claim to the popular initiative process. 9 Soft as opposed to hard sanctioning mechanisms, auch as elections or recall procedures, where the accountor (the rulers) may be chased from office (Trechsel 2010, 1059). 5

6 gathering as Saward puts it, I represent these people because they have explicitly supported my view on this issue (Saward 2009, 12). They are thus, in this definition, representatives. In Saward s view, non-elected representation can be beneficial to representative systems (Saward 2009, 9). Contrary to elected representatives, non-elected ones can, among other things, be explicitly partial, are not influenced by electoral cycles, and must put more effort to make their representative claims convincing and create an audience (Saward 2009, 8). Their presence can correct failures of representation by elected representatives, uncover ignored interests, or make new constituencies emerge. Unelected forms of representation serve as a reminder that democracy is not just about deliberation within established forums and provide a solution to avoid that those forums [ ] become sclerotic if they are not subject to pressure and renewal through outsider activism and dissent (Saward 2009, 20). Yet unelected representatives act, precisely, without the citizens control provided by elections. They are neither authorized nor accountable, which raises the question of the democratic legitimacy of their claims and actions. For Saward, the lack of institutional mechanisms can be compensated. He proposes three alternative ways to test the legitimacy of non-elected representatives claims of representation. The first way is to make sure that the claim-makers are in a structure of accountability that can potentially constrain them in ways similar to elections. For Saward, the fact that they operate in a democratic system and context already limits their claims. The second way is to wait for constituencies to emerge and to approve of the claims made by the non-elected representative a solution that requires time and deliberation. And the third way is to make sure that the claim-makers have absolutely no link with elected institution in a way that guarantees their untaintedness and the disinterested character of their claims (Saward 2009, 15 17). In sum, we shouldn t be too focused on the two traditional criteria of democratic representation, authorization and accountability, to identify democratic representatives. According to Saward, we should rather concentrate on the authenticity of the claims made; this criterion would better preserve the positive, complementing and corrective role of nonelectoral representation in democracies (Saward 2009, 21). However, Saward admits that these alternative legitimacy tests are less strong than elections, whose strength comes from the underlying values of authorization it implements (Saward 2009, 21). Moreover, considering his insistence on the fact that representative claims should be accepted by an audience (Saward 2006, 302), Saward s proposals seem to be also weak in terms of accountability: If, as he argues, representative claims participate in the construction of identity of certain constituencies and impacts the way they can act politically 6

7 (Saward 2006, 313), it seems that some institutionalized or facilitated possibility to contest some claims and punish some claim-makers would be justified. But Saward seems to have given up on the possibility to authorize and hold unelected representatives accountable and he might be right to do so for a number of claim-making actors in democratic societies. Yet the unelected representatives emerging through the facultative referendum procedure can still be formally authorized and held accountable and hence, I argue, be considered to be democratic representatives. 10 Fellow-citizens actually authorize the referendum launchers to become representatives and thus to influence policy-making on a specific issue by signing their petition. And the entirety of the citizenry can later hold them accountable by voting to support or to punish them at the stage of the popular vote. These two mechanisms are only valid for representation and action about the contested law, and for a limited period of time, namely until the popular vote has taken place. Table 1. Kinds of authorization and accountability mechanisms in a semi-direct system with facultative referendum (FR) Authorization Accountability Scope Time-frame Representative institutions (Candidates VS. Candidates) FR (Elected representatives VS. FR launchers) Regular and fair elections Gathering of signatures (Deauthorization of the representatives) (Strong); Regular and fair elections (Soft); Popular vote (Sanctioning either the representatives or the FR launchers) Competence on a bundle of issues Competence on a specific, unbundled issue Entire electoral term From the moment the FR is accepted until the popular vote Table 1 highlights the different mechanisms present in a representative democracy with a facultative referendum institution. While the representative institution only need the election to be considered democratic, i.e. to guarantee authorization and accountability, the facultative referendum adds two different mechanisms, i.e. the gathering of signatures and the popular vote, which can be used at any time during the electoral term. These two mechanisms have a simultaneous impact on the representatives and the facultative referendum launchers: The signature collection un-authorizes the elected representatives and authorizes the referendum launchers; and the popular vote sanctions either the former or the latter. 10 I would like here to thank Jean Terrier for pushing me on this point. 7

8 Compared to a pure representative political system, the one including facultative referendum procedures thus adds (1) two new mechanisms of authorization and accountability, (2) the possibility to unbundle the issues upon which representatives can act, and (3) the opportunity for unelected democratic representatives to influence policy-making. 2. No Trustees no More The additional authorization and accountability mechanisms provided by the presence of the facultative referendum have been shown to have an impact on the behavior of elected representatives. A number of empirical research on the Swiss semi-direct democracy have suggested that, because representatives want to avoid the veto of the facultative referendum, it gives them incentives to for instance include the opposition in the governing coalitions, adopt a strategy of co-operation, find compromises 11 that are acceptable to all representatives (Papadopoulos 2001, 39), or negotiate ex ante with parts of the civil society. The point I would like to make here is that the facultative referendum also impacts the way we conceive of the role of representatives in a system with facultative referendum compared to a pure representative system. As mentioned, in representative political systems, elected officials should represent their constituency during their entire term. Since Hanna Pitkin s foundational book on representation (Pitkin 1967), theorists of representation have debated about what representing exactly mean. In order to stay focused on the claim I want to make here, I take as granted the widely shared view that what democratic representatives should represent are the interests of their constituency, thus acting as substantive representatives. Their actions as representatives are thus considered valuable if they serve the interests of the (plurality of) represented (Dovi 2014). Democracy theorists have developed different conceptual models of democratic representation that representatives can adopt to act in a legitimate way. The most traditional model, which Mansbridge names promissory representation, focuses on the need for voters to keep legal or moral control over their distant representatives (Mansbridge 2003, 516). Its basic logic is the following: During the electoral campaign, candidates make promises to the voters. Once elected, they have to realize these promises during their term. If they fail to do so, the citizens can sanction them at the next election by not reelecting them. On the opposite, if they succeed, the voters can support them further by reelecting them. 11 Often referred to as consensus in the Swiss empirical literature (see e.g., Sciarini and Trechsel 1996, 30), although the term compromise seems more appropriate considering the large minorities in parliamentary votes. 8

9 Mansbridge distinguishes between different kinds of promises that candidates can make. The nature of the promise made places the representatives on the mandate-independence continuum presented at large by Pitkin (Pitkin 1967, ). If they promise to follow the constituents instructions or expressed desires (Mansbridge 2003, 516), they commit to acting as mandate, or delegate representatives. The most extreme version of the mandate conception demands that the representative acts on explicit instructions from his constituents without any exercise of discretion (Pitkin 1967, 146). A less extreme view allows the representative to act as he thinks his constituents would want, unless or until he receives instructions from them, and then he must obey (Pitkin 1967, 146). This view assumes that the represented have a rather good idea of what is good for them, and that they should be the ones deciding at least in cases of conflict between the representatives and their constituency. Other candidates can promise to further the constituency s long-run interests and the interests of the nation as a whole (Mansbridge 2003, 516). These are the kind of promises made by trustee representatives. Situated on the other side of the mandate-independence continuum, the extreme trustee must be completely free to use his own judgment without any interference from the citizens (Pitkin 1967, 146). In this view, representatives are considered to know what is in the interest of their constituency and of the national interests (Pitkin 1967, 147) better than the citizens: A good trustee is a wise judge of his or her constituents needs and how to meet them (Plotke 1997, 34, note 15). They thus can legitimately make decisions that conflict with the preferences of the citizens. Acting contrary to [the constituents ] wishes is not necessarily wrong, not necessarily bad representation or a violation of a representative s duty (Pitkin 1967, 164) in this view. The only duty for representatives in such cases is, according to Pitkin, to explain and justify their decision to go against the citizens preferences in a satisfactory way (i.e. not for private interest reasons) (Pitkin 1967, 164). All the promissory models that are found between the two extreme positions of this mandate-independence continuum are considered as defensible models of representation. Theorists then tend to favor some over others once they include additional considerations (Pitkin 1967, ). As such, deliberative democracy theorists tend to favor a model close to that of the trustee: The category of trust has proved more fruitful than that of delegate, reconfigured so it is clear that, as a trustee, the representative is obligated to keep his or her constituents in view (Urbinati and Warren 2008, 401; see also Arneson 2003, 216). One of the main deliberative democracy arguments for representative systems is indeed that representatives are more competent than citizens in making binding decisions for the 9

10 society. Some argue in this direction for rather elitist reasons: Representatives would be more competent independently of whether citizens have had an opportunity to deliberate in an appropriate, deliberative setting or not. Mansbridge expresses such a view when she argues that if citizens assemblies did binding decisions, it would probably generate costs in the legislators having less ability, expertise, and possibly even commitment to the common good than legislators chosen through election (Mansbridge 2015, 267). But most deliberative theorists argue that representatives know better precisely because of the deliberative setting in which they operate, which enables them to take various perspectives into account, to gain knowledge and information on a specific issue, and to develop considered, reasonable opinions on which to ground new legislation (Manin 1987; Cohen 1997; see also e.g., Warren and Gastil 2015, 568). Good deliberation is in their view necessary for binding decisions to have political legitimacy; yet it can only happen at the small-scale, with a limited number of participants as is the case in parliament. For deliberative democracy theorists, there are thus good reasons for citizens to trust their representatives, and to a priori accept that they do what is best for them (Pitkin 1967, 163). After all, if citizens could deliberate in the same conditions, they would arrive to similar conclusions. They can thus assume that they can approve what the representative has done (Pitkin 1967, 163; see also Hug 2008, 257), even if they themselves have neither access to all the information on an issue, nor taken part in the deliberation process. The facultative referendum institution makes the opposite assumption with regard to trust in representatives. It rejects the trustee view s fundamental assumption according to which the representatives know better : The institutionalization of a way for citizens to contest decisions made by the representatives can actually only be explained if we assume that the latter can sometimes be wrong namely, make choices that citizens judge contrary to their interests. In such cases of disagreement, it is not enough that representatives give acceptable reasons for their decision, as in the trustee model. As mentioned above, when there is a conflict about a specific law between the representatives and a minority of the population, having managed to launch a referendum, the representatives lose their authorization to make decisions about this law. Their mistake on a specific issue constitutes a good-enough reason to enable the represented to reclaim, and obtain the power to have the last word over a policy (Trechsel 2010, 1056) in order to correct the situation (Bauer and Fatke 2014, 53). This doesn t mean that unpopular policies can never be adopted; but for them to have a chance, the representatives must be able to give good reasons why they propose to undertake such a 10

11 seemingly unpopular policy (Frey and Stutzer 2013, 508) that the citizens actually accept. If they fail to convince the majority of citizens, the representatives have to see their decision rejected and they law project called an error (Sager and Bühlmann 2009, 201). In this sense, the facultative referendum changes the conception of the role of representatives by making it impossible (and undesirable) for the latter to act as trustees. The citizens are the ones who know better what their interests are: Each person is the only reliable measure of his own interest; no one else can know what it is as well as he does (Pitkin 1967, 159). The facultative referendum does not exclude that representatives might also know what the interests of the represented are. But in case they would misunderstand them, citizens must be able to signal and correct this error to control the trustee, as Paul Bauer and Matthias Fatke innocently put it (Bauer and Fatke 2014, 53) at any time. Only in this way can we avoid implementing decisions of the representatives that are not supported by the majority of the citizens, a situation deemed undesirable in the logic of the facultative referendum. This discussion thus highlights a fourth difference in the representative institutions between a pure representative system and a semi-direct one including the facultative referendum: Contrary to a system where only an elected political elite has both deliberative and decision-making power, unlike the citizens, whose formal freedom to discuss and criticize proposals and policies does not ensure that their opinions will affect legislation and policy making (Urbinati and Warren 2008, 407), the presence of the facultative referendum limits the possibility for representatives to act as trustees and enables citizens to get the last word on a specific, controversial issue. This limitation (which has, to my knowledge, never been theorized in this way) seems to induce certain questionable behaviors among representatives that have been highlighted by empirical findings about the Swiss semi-direct democracy. The presence of bottom-up direct institutions has for instance been shown to have the effect of demobilizing the representatives (Jaggi 2016 (my translation from French)). Hug argues that increasing the possibilities of sanctioning the agents reduces obviously the latter s incentives to get informed and specialized, because the efforts put into designing a law project could be annihilated by a facultative referendum (Hug 2008, 261, 260). 12 Moreover, a frequent use of facultative 12 Another explanation that I haven t so far found in the literature could be that representatives don t feel the need to specialize so much because they rely on the facultative referendum check; if this was the case, could it indicate a sense of shared responsibility of the representatives with the citizens, the latter being responsible for checking, and indeed correcting the potentially bad laws adopted by representatives? Such a view could indeed correspond to the empirical findings from Switzerland according to which the facultative referendum, despite its 11

12 referendum procedures would encourage representatives to act in a less trustworthy way (Bauer and Fatke 2014, 54). These findings seem to offer support to Philip Pettit s claim that the trust of citizens in their authorities which puts them in a vulnerable situation, as any act of trust gives an incentive to those in power to do their best (here in terms of gaining knowledge or of acting in a reliable way): In a very act whereby the trustor is put at risk, the trustee is given a motive not to let that risk materialize (Pettit 1999, 269). It also raises normative questionings: Does the facultative referendum, by excluding the possibility for representatives to act as trustees and by letting citizens reclaiming the last word on policies, impose a conception of representatives as mandates? As mentioned, in this view, the representative was sent to pursue his constituents will and not his own and loses his independence (Pitkin 1967, 146). Adrian Vatter and Deniz Danaci claim that this view indeed dominates when they write that the Swiss concept of sovereignty is that of Jean-Jacques Rousseau (Vatter and Danaci 2010, 217), namely an ideal of general will that cannot be represented (Urbinati 2006, 60), because in their view, the facultative referendum on laws demands that every law receives the confirmation of the people (Vatter and Danaci 2010, 217 (my translation from German)). This confirmation can be explicit when a facultative referendum is organized. Or it can be implicit, if we assume that laws that are not contested through a facultative referendum procedure are de facto supported by the citizens. However, the extreme mandate representative is a very objectionable conception of the representative: Representation as a contract (delegation) is in itself a denial of political representation (Urbinati 2006, 57). This model actually leaves no autonomy to representatives (Pitkin 1967, 152); insisting that the representative cannot be persistently at odds with the desires of his constituency, it goes too far in removing all his discretion to act (Pitkin 1967, 153). Representatives indeed need to have some level of independence, be it only to act upon issues on which the represented has no will (Pitkin 1967, 163). Does the facultative referendum institution thus contains this undesirable view of representation? 3. Or another kind of judged trustee? fundamental assumption that representatives can t be trusted, actually increases trust in them. In Bauer and Fatke s view, this is due to the possibility to keep their agents on a much shorter leash than in pure representative systems (Bauer and Fatke 2014, 52 53). In this sense, it could indicate that Swiss citizens consider themselves as responsible to be the watchdogs of representatives? And that the lower engagement of representatives could be compensated for by a higher engagement of the citizens? Bruno Frey and Alois Stutzer seem to suggest something similar when they write that direct democratic institutions provide incentives to voters to inform themselves about political issues, and changes their relationship to authorities and fellow citizens (Frey and Stutzer 2013, (emphasis added)). 12

13 In this still in development part, I suggest that the conception of representatives promoted by the facultative referendum does not correspond to that of the extreme mandate. Rather, it could contain a view of representation that implements a kind of trust in the political system that requires judgment on the part of citizens, and that thus answers the objection of radical democrats according to which representative governments pushes citizens to disengage from politics (Cohen and Fung 2004, 171). Several democracy theorists have recently called for enlarging our focus beyond electoral processes to think about political representation, and emphasized the need for citizens to remain active even after the election (Saward 2006; Urbinati 2006; Warren and Gastil 2015). Deliberative and participatory critiques to pure representative systems have highlighted that these systems can only be considered democratic if the citizens stay active. Joshua Cohen and Archon Fung present three of these objections to pure representative systems. First, focusing on the election moment alone would encourage citizens to excessively rely on their representatives and to renounce to act in a responsible way. Capacities of citizens may in turn atrophy, on the one hand, and on the other hand they may refrain from judging public business except under dire circumstances and then judge poorly (Cohen and Fung 2004, 171). Second, representative systems legitimized by the universal suffrage might violate equality, because it conceives of it without taking social inequalities into account. Yet deliberative and participatory democrats consider that these inequalities have an impact on the use of political power, and that they should be fought back by expanding and deepening citizen participation so that the concentration of power can be challenged (Cohen and Fung 2004, 171). Third, pure representative democracies pull the political power away from citizens. Instead of reflecting the ideal of self-government, at its best, the process reflects fair bargaining among competitive interests (Cohen and Fung 2004, 172). In Cohen and Fung s view, it should rather ensure that political arguments and appeals to interests are framed by considerations such as fairness, equality, and common advantage so that every citizen whose views do not win out can see that the decisions are supported by good reasons (Cohen and Fung 2004, 172). These objections dismiss the traditional trustee view of representation, detrimental to democracy for it conceptualizes the citizens as inactive between the elections: Once they have selected their representatives, they can rely on them to make the decisions for them, and only have to judge at the next election whether they support the decisions taken during the term. The extreme, Burkean model of trustee depends on deferential citizens who suspend judgment so that representatives might exercise their own (Warren and Gastil 2015, 564). 13

14 Although less extreme, the deliberative model of trustee mentioned in the second subsection of this paper also limits the incentives for citizens to judge the work of representatives. Because it insists that only those who took part in the deliberative process, namely the representatives, have all the information necessary to make a considered and legitimate decision, it appears to exclude the possibility for citizens who didn t take part in the deliberation to judge their representatives action. The deliberative and participatory critiques however make clear that these notions of trust should be rejected, and that the citizenry must be continuously active for elections to guarantee that representatives will act in a responsive way (Pitkin 1967, 153). Recent literature on representation has thus shifted the focus from electoral mechanisms to ongoing interactions between representatives and represented during electoral terms. As Nadia Urbinati puts it, it is desirable that the relationship between representatives and represented does not end at the moment of electoral authorization (Urbinati 2006, 58). Empowered citizens in a democracy should make ongoing judgments as voters, petitioners, advocates, deliberators, and jurors, and political institutions should be responsive to citizens judgments (Warren and Gastil 2015, 563). This is not to say that citizens should be continuously watching their representatives. On the one hand, this task would be impossible: even the most attentive citizen cannot knowledgably engage with most of the collective decisions that affect them (Warren and Gastil 2015, 562). And on the other hand, generalized distrust would undermine the citizens capacities for democratic self-government because it causes them to often withdraw from participation altogether (Warren and Gastil 2015, 563, 566). For Mark Warren and John Gastil, a certain level of trust is needed to enable citizens to play their role, namely to focus their activity on participation in specific areas or on overseeing the institutions or representatives they don t trust. Democracy thus still requires some sort of trust in (parts of) the elected and the institutions. Yet Warren and Gastil insist that this trust cannot have the desirable effect if it is a priori given to any institution or any representative. 13 Trust should, in their view, be the result of a 13 Their view here differs from that of Philipp Pettit, who, despite his insistence on citizens vigilance, promotes a view of impersonal trust according to which institutions make it possible to have an assumption of trust towards all members of the society: When we rely confidently on someone to behave in a certain way because they are independent sanctions available that support that form of behavior, then we trust that person in an impersonal fashion (Pettit 1999, 261). In Pettit s view, vigilance (a necessary aspect of the republic and a necessity for freedom as non-domination) does not contradict an attitude of confidant reliance on the powerful (Pettit 1999, 263). Individuals can at the same time trust the authorities and keep an attitude of distrust by insisting on keeping the checks and restrictions that limit their power (Pettit 1999, 264). 14

15 judgment: A decision to trust is a judgment (not a suspended judgment) (Warren and Gastil 2015, 562). With regards to their representatives, citizens should thus develop a form of selective trust enabling them to decide which representatives they should spend time and energy watching and which ones they can rely on without the extra effort (Warren and Gastil 2015, 565). These judgments are not made once and for all for the entire length of an electoral term: The context of conflicting interests continuously tests the bond between truster and trustee (Warren and Gastil 2015, 566). To offer the necessary checks, citizens trust should be well-founded (Warren and Gastil 2015, 564). For Warren and Gastil, this implies that political institutions should help citizens to make better trust judgments by creating facilitative trustee institutions (Warren and Gastil 2015, 567). Facilitative trustees facilitate, enable, and support citizens capacity for political judgment by lowing their cognitive costs (Warren and Gastil 2015, 566). Warren and Gastil suggest spreading the use of deliberative mini publics such as the British Columbia Citizens Assembly or the Oregon Citizens Initiative Review, which would serve as trusted information proxies (Warren and Gastil 2015, 568). Let s come back to the facultative referendum institution, and to explaining for now in a very sketchy way why, despite its capacity to give virtual control over every decision made by representatives, it actually seems to promote a form of selective trust in them and in the entire political system. 1. As shown in the first part of this paper, the facultative referendum provides additional means to promote interactions between the representatives and the represented during electoral terms. Its presence promotes a view of active citizens whose possibilities of influence are not limited to voting representatives in. On the opposite, citizens have an important, controlling role to play to limit the power of representatives who, as shown in the second part of this paper, may happen to make mistakes and fail to represent the interests of all the represented. This gives them an incentive to keep some minimal level of information about the work of the authorities, especially when they come to legislate on issues upon which these citizens have high stakes Citizens thus develop a form of selective trust similar to the concept offered by Warren and Gastil. However, instead of judging which representative they can 14 However, it goes further than what authors like Urbinati, Fung and Cohen, or Warren and Gastil argue for by enabling citizens to also make binding decisions upon laws (see e.g., Urbinati 2006, 58; Cohen and Fung 2004, 172). This point needs some further considerations. 15

16 trust or not, they judge what issue they have to oversee or not. They don t have to engage with every decision of the parliament, but only with those that affect them. In other words, they can decide to trust the parliament for all the topics in which they have no or low stakes as long as the facultative referendum is not triggered. 3. Yet if representatives happen to adopt a law that goes against the interests of a part of the constituency, the latter can launch a referendum. In this case, the entire citizenry has to make a trust judgment about this specific issue. The fact that a part of the citizens claim to have their interests hurt by a law demands that all the citizens judge the issue for themselves and decide whether to support the elected representatives or the facultative referendum launchers. 4. In this way, the facultative referendum launchers appear to play the role of facilitative trustee. As long as there is no contestation through a facultative referendum, citizens can trust their representative institutions. But when the nonelected representatives launching the referendum highlight the fact that representatives might fail to represent them properly on a specific issue, they indicate to the entire citizenry that they should remove their trust in the representatives on that issue and engage with this topic. They facilitate the trust judgments of citizens by informing them that the quality of representation is questionable, thus functioning as information proxies (Warren and Gastil 2015, 658). Just as members of mini-publics, they can be trusted because they can be assumed to have impartial, in the sense of disinterested, motivations and of being (minimally) competent, because they are sufficiently interested in politics to launch a facultative referendum (Warren and Gastil 2015, 568) (but also, maybe, because it is their interests that are at stake). Moreover, their use of the facultative referendum increases the accessibility to information about the law contested, as both the launcher and the representative campaign before the popular vote in an attempt to justify their position to the entire constituency (Setälä 2006). For these reasons, which still need to be reflected upon, I posit that the semi-direct political system with facultative referendum rights promotes selective trust and an engaged citizenry. Before I close this paper with some remaining challenges and questions, I want to come back to the implications of the facultative referendum for the role of representatives. I actually mentioned earlier the empirical findings according to which representatives in the Swiss semidirect system don t specialize as much as representatives in pure representative systems. Yet I 16

17 would like to suggest that, if they don t get the same level of knowledge about the issues at stake, they might get more of another kind of knowledge about citizens expectations and interests. In a system with the facultative referendum, elected representatives actually have a strong incentive to get to know what are the opinions shared among the citizenry. Just as elected representatives in pure representative systems want to be reelected to keep their prerogatives in the political decision-making process, representatives in mixed-systems actually want to avoid the challenge points of referendums, and so to keep their authorization to act on all the issues of which they are in charge, in addition to being reelected. In a semi-direct democracy, the aims of representatives could thus be both to be reelected and also to avoid popular scrutiny (Hug 2008, 261) that would withdraw their authority to act on a certain topic. At the very least, they have an incentive to collude with the actors being able to trigger a referendum, if possible, to avoid popular scrutiny (Hug 2008, 261). In this sense, the model of anticipatory representation developed by Mansbridge 15 seems to describe their role in a very relevant way. Contrary to the promissory model, which focuses on the realization of the promises made by the representatives at the time of their election, the anticipatory model highlights the importance of the potential reelection moment in they way representatives choose to act. Because of their desire for reelection, anticipatory representatives would actually act so as to please future voters namely voters at the time of the next election (Mansbridge 2003, 518, 517). Their actions would be guided by their beliefs [ ] about the future preferences of the voter (Mansbridge 2003, 517). Representatives would thus have an incentive to be in continuous communication and interaction with the voters in order to learn about and keep track of their opinions. In a prudential move to realize their goal of reelection, they would give attention to underlying interests as well as present preferences (Mansbridge 2003, ). According to Mansbridge, anticipatory representatives play a role similar to that of entrepreneurs anticipating the demands of clients : They are active, in searching out and sometimes creating new preferences (Mansbridge 2003, 518) A model to which Pitkin shortly points at when she writes that legislators often pattern their actions not on what their constituents ought to want but on what they anticipate their constituents will want [ ]. This is natural; the legislators want to be reelected (Pitkin 1967, 164). 16 The analogy of clients is quite problematic with regards to clientelism, which may sometimes be functional for democracy in that it provides a means for parties to mobilize support and for candidates to build resilient constituencies but comes with many dangers for democratic regimes (Brun and Diamond 2014, ix); Mansbridge doesn t seem to consider this in her article, but it could deserve additional considerations. 17