• 1 Introduction

  A dominant criticism in the democratic innovations literature with respect to referendums is their tendency to promote a single proposal that voters can only accept or reject. The inability to make compromises can aggravate rather than solve policy conflicts (e.g. Altman, 2018; Morel, 2018a). It also provides the author of the ballot proposal and the actor initiating the referendum – which may be the same or two different actors – with significant influence over referendum process and outcome (e.g. Hug & Tsebelis, 2002; Jäske & Setälä, 2019; Wagenaar, 2021). There are, however, ways in which referendum processes can be adapted to comprise improved or additional policy proposals, which are more in line with societal preferences than the sometimes narrow interests of the initiator or the author of the initial proposal. This article reviews innovations to referendum processes that deviate from the traditional binary referendum on a predetermined proposal. Such innovations have been proposed by various academics and in diverse contexts. This article presents them alongside one another and reviews their applicability for referendums in the low countries, where referendums are being held predominantly at the local level and where offering a real choice to voters and interpreting how to adapt rejected legislation regularly poses challenges to local decision-making. At the same time, the decentralised level lends itself well to experimentation with alternative referendum procedures in diverse contexts and on diverse topics, and the proximity to citizens facilitates following up on how legislators deal with citizen input (Verhulst & Nijeboer, 2007). The proposed innovations in referendum initiation and ballot formulation described in this article are compared with empirical examples as well as suggestions emerging from society in the low countries aimed at increasing the constructive value of referendum processes as a tool of citizen participation.
  Section 2 introduces two often-used dimensions to classify (1) who triggers – or in other words initiates – a referendum and (2) what the relationship of the triggering actor is to the author of the ballot proposal on which the referendum takes place – and therefore what the referendum vote intends to accomplish. Referendums thus differ according to their initiation (top-down or bottom-up) and their intention (proactive or reactive). Section 3 summarises the limitations of the traditional one-proposal binary referendum without provisions for societal input or amendment, signalling the relevance of analysing alternative ways to design referendum processes. Section 4 reviews the literature on proposed and implemented innovations that affect the options presented on the ballot. This article explores alternative referendum procedures that provide for feedback procedures, additions to the ballot or direct changes to the referendum proposal by actors other than the referendum initiator or author of the original proposal. The innovative variants discussed all address, in different ways, the common criticism that referendums enable citizens solely to approve or reject a predetermined policy proposal. Section 5 then considers the potential for applying the reviewed innovations in the low countries. After briefly introducing referendum experience and legislation in the three countries, this section discusses the potential for referendum innovation, describing concrete examples wherever applicable and discussing suggestions that have emerged in society to inspire innovation. The main focus is on the Netherlands, but the suggestions are largely applicable to the similar local referendums contexts in Belgium and Luxembourg. Section 6 concludes with several recommendations for practical experimentation.

• 2 Referendum Characteristics: Initiation, Intention and Ballot Authorship

  The referendum literature distinguishes two main dimensions for classifying referendums: who initiates the referendum and what its intention is: to propose new legislation or to correct a legislative decision. Referendums are traditionally binary in nature. As a consequence, regardless of who triggered the referendum, the single referendum proposal author has a great deal of agenda-setting power (Altman, 2011; Hug & Tsebelis, 2002; Suksi, 1993; Wagenaar & Hendriks, 2021). Referendum triggering under traditional models impedes amendments being made to the proposal. Before looking into possible innovations, the next two subsections further explain the dimensions of referendum initiation and intention.

  2.1 Referendum Initiation: Top-Down or Bottom-Up

  Most referendums held around the world are facultative, meaning that they are not required by the constitution¹x Facultative referendums contrast with mandatory referendums. In some countries, such as Ireland, the constitution stipulates that the electorate must approve any changes to the constitution in a referendum. The referendum is thus said to be triggered automatically, contrary to the facultative, i.e. voluntary, triggering of a referendum by particular actors (e.g. Breuer, 2008). Since mandatory triggering does not actively involve referendum initiators, such referendums fall outside the scope of the theoretical framework for this article. but are voluntarily triggered by a particular actor (Altman, 2011; Suksi, 1993). General requirements for holding referendums can be regulated in legislation, but individual referendums require active triggering, as opposed to the mandatory category (Suksi, 1993: 28). Legislation specifies whether a referendum can be triggered by, for example, executives or a specified number of parliamentarians or citizens. Triggering by legislatives or executives is referred to as top-down (Papadopoulos, 1995). Particularly those triggered by executives (a head of state, prime minister or government) are also referred to as plebiscites in the referendum literature and are often viewed critically as a way of legitimising a government-preferred proposal rather than genuinely extending citizen influence over policymaking (Altman, 2011; Kaufmann & Waters, 2004; Morel, 2018b). When citizens can qualify a policy topic for a referendum vote by collecting a predefined number of supporting signatures, this is referred to as bottom-up triggering (Papadopoulos, 1995). The process is commonly spurred by an existing or newly formed interest group (Suksi, 1993). The literature is divided over the question of whether referendums triggered by an institutional minority (for example one third of parliamentarians) should be classified as bottom-up (e.g. Morel, 2018b) or top-down (e.g. Stefanini, 2018). If there are no explicit legal regulations specifying the conditions to initiate a referendum, referendums can still be triggered by a majority of parliamentarians by passing ad hoc legislation for a specific referendum, in a so-called non-pre-regulated referendum (Suksi, 1993).

  2.2 Referendum Intention: Proactive or Reactive

  The intention of the referendum can be either to change the status quo by advancing new legislation, i.e. to be proactive, or to sustain the status quo by correcting new legislation that seeks to change it, i.e. to be reactive (Altman, 2011).²x Others have used different terminologies for proactive or reactive characteristics, such as propositive versus non-propositive (Morel 2018b) or decision-promoting versus decision-controlling (Uleri, 1996). This dimension relates to the question of whether the referendum initiator also authors the referendum proposal. In proactive referendums, this is the case, whereas in reactive referendums initiators trigger a referendum on a proposal authored by another actor. In the case of bottom-up referendums, citizens can advance a citizens’ initiative to propose new legislation (proactive) or can collect signatures to correct a decision by policymakers (reactive). Suksi (1993: 30) also uses the terms active and passive participation to denote whether citizens can or cannot influence which issues or proposals are submitted to the referendum vote. Top-down referendums are predominantly proactive, as there is no logic in a corrective referendum if both the triggering actor and the author of the legislative proposal are the same entity (commonly the parliamentary majority), thus proposing to correct one’s own proposal. Reactive top-down triggered referendums may occur when executives can trigger referendums, such as in presidential systems.³x For example, in various Latin American countries, both the president and parliament have legislative as well as referendum-triggering powers, which can result in a top-down triggered, reactive referendum on a proposal formulated by the other entity (Breuer, 2008). These less common referendum types are not the main focus of this article but could arguably be subject to similar innovations as those discussed for bottom-up triggered reactive referendums.

  The dimensions of initiation and intention, summarised in Table 1, serve as a basis for clustering referendums with a similar design, notwithstanding individual differences such as signature requirements and topical exemptions. These two dimensions form a useful basis for an exploration of innovations aimed at more constructive referendum processes, as the characteristics of initiation and intention may encourage or impede particular types of innovations.

  Table 1 Dimensions of referendum initiation and intention

  		Referendum initiation
  		Top-down – triggered by legislators or executives	Bottom-up – triggered by citizen signatures
  Referendum intention	Proactive – triggering actor authored the ballot proposal	Propositive referendum – referendum on a new legislative proposal authored by legislators	Citizens’ initiative – referendum on a new legislative proposal authored by citizens
  	Reactive – triggering actor reacts to ballot proposal authored by another entity	Corrective referendum – referendum to correct a legislative proposal authored by legislators	Corrective referendum – referendum to correct a legislative proposal authored by legislators

• 3 Ballot Limitations of Traditional Binary Referendums

  Referendums have various advantages over other democratic innovations, such as their efficiency and inclusiveness (Michels, 2011; Taillon, 2018). Research consistently shows a high level of support for referendums as well as more diversity in regard to participating citizens compared with other instruments of citizen participation (e.g. Altman, 2011; Gastil & Richards, 2013; Jacobs, 2018; Mendelsohn & Parkin, 2001; Taillon, 2018). Opponents and critics of referendums have, however, pointed to various limitations of the referendum instrument, including the bluntness of binary choice: citizens are given decision-making influence but only on a specific proposal that they can accept or reject (e.g. Bochsler, 2010; Lupia & Johnston, 2001; Mac Ginty, 2003; Wagenaar, 2019). A yes/no vote discourages reasoned deliberation (Rein, 2008) and impedes democratic co-creation. This may force voters to express their preference for the lesser of two evils as opposed to a genuinely preferred policy alternative (Morel, 2018a). This lack of more nuanced options may oversimplify policy issues (Setälä, 1999; Taillon, 2018), polarise the electorate (Altman, 2018; Parkinson, 2001) and distort referendum outcomes (Şen, 2015). Because of their inability to incorporate compromises or concessions, binary referendums may aggravate conflicts rather than resolving them (Altman, 2011; Morel, 2018a). In ordinary top-down referendums, citizens have influence neither over the topic of the vote nor over the specific proposal to be voted on. In conventional reactive bottom-up referendums, initiators can only advocate rejection of a legislative proposal, with no power to propose improvements or alternative policy directions. In a proactive bottom-up referendum (a citizens’ initiative), a small group of initiators does have influence over the proposal put forward, but segments of society may consider an alternative proposal more suitable. Although referendum processes are not designed to involve all citizens in elaborating on policy decisions together (Morel, 2018a), there are ways to provide for amendments and additional ballot proposals that can bring referendum ballots and results more in line with societal preferences. The next section discusses innovations aimed at generating such more constructive referendum ballots.

• 4 Innovative Developments: Extending or Amending Referendum Ballots

  Research and experimentation are increasingly devoted to the question of whether referendums can be innovated to improve ballot choice for voters. This section reports on a number of variations on the standard one-proposal referendum in which either the legislature or a societal referendum initiator is the sole and final author of the ballot proposal. An ‘innovation’ is defined as the inclusion of one or several procedural steps that affect the content of the proposal or proposals offered on the referendum ballot. Some of the innovative referendum procedures discussed have been in use for several decades – notably in referendum-minded democracies – but nonetheless deserve attention as their use is still far from standard and they might serve as inspiration for application elsewhere.
  Broadly speaking, there are three non-mutually exclusive procedures through which referendum ballots can be innovated: (1) delegating the formulation of all or some ballot proposals to increase representativeness, (2) incorporating provisions for amending an initial proposal, resulting in a binary referendum on an improved proposal and (3) extending ballot choice by adding additional alternatives to the ballot. In the remainder of this section, various innovations applying such procedures, either empirically practised or theoretically proposed, will be discussed according to the referendum initiator (top-down or bottom-up) and, for bottom-up referendums, their intention (proactive/reactive). Their procedural steps are visualised in Figures 1-3 alongside the traditional prototype described in Section 2 and Table 1.

  4.1 Innovations in Propositive Referendums (Top-Down with Proactive Intentions)

  The main criticism of top-down referendum ballots is their strategic use by elites, who can manipulate the ballot proposal to affect the referendum outcome in their favour (e.g. Hug & Tsebelis, 2002; Setälä, 1999). This subsection considers innovations that have been identified in the academic literature to bring ballot choice for top-down triggered, propositive referendums more in line with societal preferences (see Figure 1).

  Innovations in ballot proposals for propositive referendums
  Note: Lined circles denote a delegation of the formulation of ballot proposals to actors other than the referendum initiator. Double-lined circles denote an extension of ballot choice.

  4.1.1 Ballot Proposal Designed by Citizens’ Assembly

  A first type of innovation aims to bring the content of a single proposal submitted to the referendum ballot in line with societal preferences by involving a representative group of citizens in its formulation. In recent years, various countries have experimented with referendum proposals formulated by randomly selected citizens’ assemblies representative of the electorate. Notable examples include referendums on abortion legislation in Ireland (Farrell, Suiter & Harris, 2019), on constitutional change in Iceland (Landemore, 2015) and on the electoral system in the Canadian state of British Columbia (Warren & Pearse, 2008). The triggering of such referendums can be considered top-down, as legislators or executives explicitly delegate decision-making on a particular policy issue to citizens.⁴x In the case of Ireland, referendums on constitutional changes are mandatory. In order to revoke the prohibition of abortion from the constitution and to allow parliament to legislate on abortion, popular approval in a referendum was obligatory. Since the citizens’ assembly could, however, not formally pass legislation, the parliament acted as an intermediator by formally proposing to change the constitution and to legislate for abortion in line with the assembly’s proposal. Authorising a citizens’ assembly to draft a referendum proposal for approval by the electorate has also been suggested by Gastil and Richards (2013), particularly for issues that might pose conflicts of interest, such as electoral rules and campaign reform, or regulations on political speech (see also Ferejohn, 2008). Whereas the foregoing empirical examples required that parliament formally triggers a referendum on the proposal that has been developed by the citizens’ assembly, Gastil and Richards (2013) propose a ‘silent approval’ process. This would entail a procedure in which the draft legislation formulated by the assembly is automatically submitted to a referendum vote unless legislators explicitly block this through a majority vote. Such a procedure links citizen authoring of a ballot proposal directly to popular approval, with more limited scope for parliamentary intervention.

  4.1.2 Multiple Ballot Proposals Formulated by a Non-legislative Entity

  Rather than focusing on the content of a single referendum proposal, this innovation diversifies the choice presented on the ballot, thereby directly addressing the dominant limitation of conventional referendums in offering a single proposal for approval or rejection. Referendums can be designed such that they offer multiple proposals: several variations on a policy or alternative scenarios for dealing with a policy issue. Such referendums have been hailed in the literature under various names, such as multiple option referendums (Gallagher, 2014; Orr, 2001), multi-choice referendums (Barber, 1984) and multi-option referendums (Mitchell, 1992; Tierney, 2013; Wagenaar, 2021). Referendums with multiple ballot options can encourage constructive voting (Mendelsohn & Parkin, 2001) and reduce elite control over the referendum process (Lupia & Johnston, 2001; Tsebelis, 2018) while maintaining the aggregative benefits of referendums, as long as provisions are in place to ensure an unequivocal winning option (Wagenaar, 2021). In top-down triggered multi-option referendums, legislators initiate the referendum but usually delegate the formulation of the ballot options. The ballot options can be designed, for example, by different political parties or coalitions, as has been practised in Sweden (Bjørklund, 1982; Ruin, 1982; Suksi, 1993). Alternatively, the task can be delegated to a committee of experts, which can either be an existing expert committee such as an electoral commission, as practised in Jersey (Wagenaar & Hendriks, 2021) or a specially instituted committee, as in New Zealand (Levine & Roberts, 1993; Nagel, 1994). A designated committee can also be tasked to select alternatives from societal inputs – rather than drafting alternatives themselves – as was the case in the 2015 flag referendum in New Zealand, where ballot options were selected from over 10,000 designs submitted by citizens (Tudor, 2016). Instead of a committee of experts, the task of exploring and selecting multiple policy proposals for the ballot could also be delegated to a representative citizens’ assembly, as suggested by Wagenaar and Hendriks (2021). The procedure, which has not been practised to date, would be an adaptation of the previously discussed innovation, except that instead of seeking consensus on a single proposal, the assembly would formulate multiple proposals and delegate the final decision to the electorate.

  4.1.3 Preferendum with Diversified Ballot Authors

  The term preferendum is used ambiguously in the literature, sometimes as a synonym for any referendum with multiple ballot alternatives, although mostly referring to a specific variant of it, in which voters decide on the options using preferential voting procedures. It was initially envisioned as a consensual voting method using the Borda count system, which allocates points to different ballot alternatives depending on the rank they receive from voters (Akkerman, 2004; Emerson, 1993, 2021; Morison & Newman, 2001; Orr, 2001). Others have proposed applying preferential voting methods other than points systems (Lundberg, 2007; Mackerras, 1994; McLean et al., 2003; O’Flynn & Levy, 2020). The original intention of the preferendum method is that diverse entities propose the various alternatives, thus diversifying the ballot drafting process compared with other types of multi-option referendums (Morison & Newman, 2001). A similar idea is coined by Tsebelis (2018), who proposes that after a referendum has been triggered – either top-down or bottom-up – other actors should be able to propose alternative proposals, to which lower signature requirements apply. This would encourage the original initiators to submit a well-supported proposal as opposed to a highly ideological plan. Such diversification could help to reduce the manipulation of referendum procedures by political or societal elites.

  4.2 Innovations in Bottom-Up Initiated Referendums

  The previous subsection discussed suggestions to innovate ballot formulation procedures for referendums triggered by a parliamentary majority or the executive. This subsection describes innovations that can be applied to referendums triggered by citizens. Some innovations bear similarity to those described previously, but an important difference between bottom-up and top-down triggered referendums is the role of the initiators. Whereas parliamentarians have more leeway to trigger ad hoc referendums and then delegate ballot formulation to, for example, a citizens’ assembly or an expert committee, citizen initiators tend to trigger a referendum either on a specific proposal of their own (proactive) or on parliamentary legislation that they contest (reactive). The steps that can be incorporated to innovate the resulting referendum ballot therefore differ from those identified previously. Rather than formulating it from scratch, the focus is on supplementing or improving the referendum proposal that formed the basis for triggering the referendum.

  4.2.1 Innovations in Citizens’ Initiatives (Bottom-Up with Proactive Intentions)

  The main criticism on the bottom-up, proactive practice is that initiatives benefit mostly well-organised societal groups or individuals with financial and political resources. The initiative process (from drafting the initiative to qualifying it for the ballot and campaigning for it) can be strategically abused to serve special interests (Altman, 2014; Broder, 2000; Schacter, 1995; Staszewski, 2003). Particularly in the US, the literature tends to be critical of what is called the ‘initiative industry’ (Altman, 2018; Magleby, 1995). For example, in California, referendums have been triggered by billionaires and powerful organisations such as the Realtors Association, who have spent as much as 2 million US dollars to qualify initiatives for the ballot. The money is spent, for instance, on lawyers and on paid signature collectors (McCarthy Carino, 2018).The academic literature also points towards increasing professionalisation in signature collection and strategic initiative drafting (Broder 2000; Magleby, 1995; Staszewski, 2003), and financial and organisational resources are said to largely influence which initiatives qualify for the ballot and succeed (Jäske & Setälä, 2019). It is therefore questionable whether initiatives serve the interests of the electorate as a whole (Magleby, 1995; Staszewski, 2003). Smith (2007: 265) describes how the initiative process tends to promote ‘quick, unthinking voter signatures rather than sophisticated public education and discussion’. Voters may favour a change in legislation but might prefer a different kind of policy amendment than the one proposed by the initiators. Under the traditional initiative model, they are unable to affect the content of the initiative (Magleby, 1995). Five innovations are discussed in which additional procedural steps involve either legislators or citizens in suggesting amendments or alternatives to the initial ballot proposal (see Figure 2).

  Innovations in ballot proposals for citizens’ initiatives
  Note: Dashed circles denote room for amending an initial proposal. Lined circles denote a delegation of the formulation of alternative ballot proposals to actors other than the referendum initiator. Double-lined circles denote an extension of ballot choice.

  4.2.1.1 Political Review Between Initiation and Final Submission

  Although Germany has very limited provisions for national-level referendums, binding referendums are possible at subnational levels in all states, including both proactive and reactive bottom-up referendums (Geissel, 2016). Proactive citizens’ initiatives can occur according to two models, a two-tier and a three-tier model (Kampwirth, 2004; Rohner, 2011), depending on the state. The three-tier model provides for mid-process amendments by the initiators. After submitting the initiative with the required number of signatures, the eligibility of the initiative is determined by the competent authority, which is usually the state ministry, government or parliament, or, in exceptional cases, the state-level constitutional court (Rohner, 2011). If eligible, the state-level parliament debates the initiative and decides to either accept it, in modified or unmodified form, or to reject it.⁵x This step is skipped under the two-tier process. Initiators can proceed to the second stage, collecting the additional signatures required to qualify the proposal for the ballot as soon as the initial request is declared admissible. This procedure is similar to bottom-up referendum procedures in, for example, the Netherlands, where the signature collection process is split into two phases, allowing for a determination of the admissibility of the referendum proposal before all signatures are collected. Under such a model, there is no possibility for the legislative to adopt the initiative or suggest amendments. The intention of this additional step in the process is to avoid referendums on initiatives that enjoy support from the legislature and could thus be introduced without requiring popular approval. In the case of a rejection or of modifications not accepted by the initiators, initiators can decide to qualify the amended proposal for the ballot by meeting a higher signature threshold (Rohner, 2011).⁶x This distinguishes the German three-tier model from so-called ‘referendum motions’, which are used, for instance, at local levels in Finland, Norway and Iceland. Referendum motions obligate a municipal council to consider organising a referendum on a policy topic if proposed by a specified number of citizens, but initiators cannot force the vote if legislators decide against it. They are also referred to as ‘soft’ forms of direct democracy (Jäske, 2017). Initiators are allowed to amend their proposal in response to the reasoning provided by parliament, and the state parliament can also submit a counter-proposal (see next innovation). The ultimate decision rests with the electorate.

  4.2.1.2 Political Counter-Proposal to a Citizens’ Initiative

  Once a citizens’ initiative qualifies for the ballot, some jurisdictions allow for the formulation of a political counter-proposal. This is relatively common practice in various German states (under both the two-tier and the three-tier models, see Rohner, 2011), in Switzerland (both at federal and at state levels) and in Liechtenstein (Hendriks et al., 2020). The counter-proposal is usually formulated by a parliamentary majority, but regulations can also attribute this right to minorities, such as a two fifths minority of parliamentarians in Uruguay (Wagenaar & Hendriks, 2021) or a parliamentary minority or upper chamber in Slovenia (Nikolenyi, 2011). Both proposals are then included on the ballot, providing the electorate with a choice between the status quo, revisions proposed by the societal initiators and revisions proposed by parliament. In case of counter-proposals, it is important that voters can accept both proposals (the so-called ‘double yes’ provision) and highly preferable that they can also indicate their relative preference in a deciding question. This deciding question (termed ‘Stichfrage’ in Swiss and German referendum terminology) pits the change proposals directly against each other in a separate question, allowing all voters, including those approving or rejecting both proposals, to indicate their relative preference. Double-yes and deciding-question provisions prevent strategic use of counter-proposals to split votes over two proposals in order to favour the status quo option (Rohner, 2011; Wagenaar, 2021). Whether legislators initiate a counter-proposal depends on various factors, including the perceived relevance of the topic, the numerical thresholds and the ideological positioning of the initiative vis-à-vis the median voter and legislators (Altman, 2014, 2018). In Switzerland, initiators may withdraw their proposal in response to a political counter-proposal, resulting in a referendum ballot posing a binary choice between the legislative counter-proposal and the status quo (Wagenaar, 2021).

  4.2.1.3 Societal Inputs Between Initiation and Voting

  Staszewski (2003) suggests the incorporation of a deliberation phase between the appeal for a citizens’ initiative and the final approval of the initiative for the ballot. His proposal is situated in the US context, where in most states citizens can collect signatures for an initiative and where judicial review of initiatives takes place prior to the vote. In this process, there are no formal requirements for initiators to engage with citizens and other interested parties to collect alternative ideas or make compromises. This promotes lobbying for the initiators’ most favoured outcome, rather than investing efforts to improve the policy proposal and align it with societal interests. Staszewski proposes to introduce a mandatory phase between signature collection and judicial review in which opponents, elected officials and interested citizens can submit comments on the content of the initiative. Initiators would be required to respond to the comments and suggestions in an amended proposal, accompanied by a written statement, or can decide to withdraw their initiative in response to criticisms. Judicial review then assesses whether initiators engaged reasonably with submitted comments. Jurors would be authorised to invalidate the proposal if important aspects of the issue are ignored, arguments run counter to evidence or public comments are not convincingly addressed. The new procedure would hold initiators accountable for the legislative process, increase societal deliberation and limit the scope for utilising the initiative process for self-interested purposes. Smith (2007: 301) agrees that requiring initiators to consult public officials, experts and interested parties can improve ballot drafting, decrease unforeseen consequences and create a ‘deliberative check’ on the initiators by forcing them to take societal perspectives into account and to respond to criticisms. He favours the consultation procedure proposed by Staszewski but prefers the involvement of interest group representatives over direct input by voters and proposes to require a mandatory public hearing on the ballot proposal.

  4.2.1.4 Citizens’ Counter-Proposal to a Citizens’ Initiative

  In a variation on the political counter-proposal discussed in Section 4.2.1.2, alternative groups of civic initiators could propose a counter-proposal to an initiative that has qualified for the ballot. Whereas in the previous innovation, a single proposal is considered for the referendum ballot and its initiators remain in charge of adapting it according to received inputs, the citizens’ counter-proposal allows for direct qualification of a competing ballot alternative. This procedure is regulated for in Uruguay, where initiators can propose a counter-proposal in response to another citizens’ initiative (Wagenaar & Hendriks, 2021). The ballot thus consists solely of a variety of citizens’ proposals competing against the status quo. This procedure is also advocated for by Tsebelis (2018), as discussed in Section 4.1.3. He considers diversified ballot option formulation desirable for both top-down and bottom-up triggered referendums. In his view, signature requirements ought to be lower for subsequent proposals, which differs from the current regulations in Uruguay, where requirements are the same for all ballot proposals. Altman (2018: chapter 7) suggests delegating the formulation of the counter-proposal to a deliberative citizens’ assembly of around 23 allotted members under compulsory and paid participation. Their counter-proposal would be voted on by the electorate as part of a two-stage referendum procedure.⁷x The staging would entail a first round of voting on whether to retain or change the status quo, in case of majority support for change, a second stage in which the alternative proposals compete for support. Random selection and internal deliberation would avoid the counter-proposal from being seized by those with private interests.

  4.2.1.5 Citizens’ Assembly between Drafting and Signature Collection

  Also in response to criticisms on the utilisation of initiatives by special interest groups, Ferejohn (2008) proposes to institute a citizens’ assembly each time an initiative is fielded. The assembly would take time to get acquainted with the issue, deliberate the initiative and propose amendments. The amended initiative is expected to be closer to the preferences of the median voter than the original. Ferejohn further expects a Sword of Damocles effect: since initiators are aware that the proposed content and language of their initiative will be reviewed by a representative assembly, they would be less likely to frame it in terms of narrow special interests. Altman (2018), however, is critical, noting that no matter how outstanding the deliberation and the resulting amended proposal, altering a citizens’ initiative betrays those citizens whose signatures qualified the initiative for the ballot. This could be forestalled by not collecting signatures until after the citizens’ assembly has taken place, as suggested by Gastil and Richards (2013). They propose a Design Panel, which would be instituted – and paid for – by the initiative petitioner on a voluntary basis after drafting the initiative but before submitting it to signature collection. The representative panel would convene for five days to discuss the initiative and make suggestions to improve or reject it. If the initiators accept the proposed amendments, the signature requirements to qualify the proposal for the ballot are substantially lowered.

  4.2.2 Innovations in Corrective Referendums (Bottom-Up with Reactive Intentions)

  Binary referendums on approving or rejecting challenged legislation can polarise the electorate and aggravate conflict within society and between citizens and politicians (Morel, 2018a). Majority rejection in corrective referendums can, moreover, generate uncertainty about whether to abolish legislation altogether or pursue an alternative (Taillon, 2018; Wagenaar, 2021). In response to the inability to make compromises in traditional corrective referendum procedures, two types of innovations seek to introduce opportunities for referendum initiators to more concretely address the perceived legislative obstacles, either by pointing out specific parts of legislation to be amended or by formulating a concrete policy alternative (see Figure 3).

  Innovations in ballot proposals for corrective referendums
  Note: Dashed circles denote room for amending an initial proposal. Lined circles denote a delegation of the formulation of alternative ballot proposals to actors other than the author of the initial proposal. Double-lined circles denote an extension of ballot choice.

  4.2.2.1 Semi-propositive Referendum: Rejecting or Improving Specific Parts of Legislation

  In abrogative referendums, used in Italy and Malta, active legislation can be submitted to a repeal procedure in which the initiators can indicate whether the legislation ought to be revoked as a whole or in part (Capretti, 2002; Uleri, 2002). Gallagher and Uleri (1996: 109-110) refer to the Italian abrogative referendum as ‘quasi-propositive’, since these procedures also allow for the correction of specific parts of legislation, rather than revoking legislation as a whole. Morel (2018b: 31) similarly distinguishes what he terms ‘semi-propositive’ referendums as an intermediate form. In his view, referendum initiators should be able to propose a replacement policy, which must be approved by another entity, usually the parliament, before being submitted to voters. Capretti (2002) has suggested a similar application to the Italian context of abrogative referendums: rather than ‘creatively deleting’ parts of legislation, citizens should also be able to suggest provisions to replace these parts. These suggestions all address the limitation inherent in traditional corrective referendums that legislation can only be rejected in its entirety, regardless of possible alterations considered sufficient by the referendum initiators.

  4.2.2.2 Citizens’ Counter-Proposal to New Legislation

  Rather than indicating specific parts of legislation for repeal or alteration, a citizens’ counter-proposal could be submitted by the initiators of a corrective referendum in which they detail a fully developed alternative policy proposal. Diverting from the semi-propositive referendum promoted by Morel, citizen signatures qualify the proposal for the ballot without requiring legislators’ support. Moreover, both proposals are voted on by the electorate. Several Swiss cantons have experience with such a citizens’ counter-proposal (Lagerspetz, 2016), which has also been termed the ‘constructive referendum’ (Glaser et al., 2016). As in the semi-propositive referendum, a citizens’ counter-proposal empowers citizens to engage constructively with the referendum topic in what is otherwise a mere reactive process (Wagenaar & Hendriks, 2021). In some cases, multiple citizens’ counter-proposals can be added to the ballot, as long as each meets the signature requirements, as has been practised in Zurich and Nidwalden (Glaser et al., 2016). International guidelines for good practice stipulate that if citizens can self-organise to draft a counter-proposal, it is desirable that they may submit the proposal in general terms as opposed to a fully specified draft. This encourages citizens without extensive legal skills to participate in the process (Venice Commission, 2007: 19). Nonetheless, similar to the issues raised with citizens’ initiatives in Section 4.2.1, the citizens’ counter-proposal can be vulnerable to hijacking by well-organised special interests. Altman’s proposal for the formulation of a citizens’ counter-proposal by a deliberative citizens’ assembly, as discussed in Section 4.2.1.4 for proactive citizens’ initiatives, in his view thus also applies to counter-proposals for reactive corrective referendums. Similarly, both McKay (2018) and Wagenaar and Hendriks (2021) suggested the implementation of deliberative citizens’ assemblies to formulate a counter-proposal in bottom-up reactive referendum processes.

• 5 Scope for Referendum Innovation in the Low Countries

  Belgium and the Netherlands are among a worldwide group of under 40 countries without national-level provisions for referendums (Morel, 2018b). Parliamentary opinion on the introduction of national-level referendum legislation in the Netherlands has been rather adamant over past decades. Despite various parliamentary commissions advising their introduction, and occasional parliamentary attempts to introduce referendum legislation, parliamentary majorities have rejected most draft laws (Hendriks et al., 2017). Legislation for a non-binding reactive referendum was briefly in place between 2015 and 2018, before being abolished after two referendum experiences (Van der Meer et al., 2020). In 2018, an advisory state commission on the parliamentary system recommended, among other things, the institution of a binding referendum (Staatscommissie parlementair stelsel, 2018: 135-154). Recent attempts to introduce binding referendum legislation at national and/or sub-national levels were, however, rejected by the Dutch parliament in July 2022.⁸x Parliamentary vote: www.tweedekamer.nl/kamerstukken/stemmingsuitslagen/detail?id=2022P13194&did=2022P13194. Similarly in Belgium, federal referendum legislation appears to be a distant possibility (Caluwaerts & Reuchamps, 2020). Belgium last experienced a national-level referendum in 1950 on the post-war return of the king.
  Both the Netherlands and Luxembourg held ad hoc top-down triggered referendums in 2005 on the EU constitution. In contrast to the Netherlands and Belgium, the Luxembourgian constitution explicitly legislates for a top-down triggered referendum. In 2015, the parliament initiated three national-level constitutional referendums – on lowering the voting age, voting rights for foreigners and term limits for government ministers. Back in 1919, a multi-option referendum was held in Luxembourg on the head of state, offering voters four options: maintain the current grand duchess; elect a new duchess under the existing system; retain the monarchy but replace the dynasty; or switch to a republic (Wagenaar, 2020).
  At the local level, the Netherlands, Belgium and Luxembourg have similar provisions for non-binding referendums, which can be triggered either by citizens or politicians (Scarrow, 2001). Their use is most frequent in the Netherlands, with over 200 referendum experiences, although scattered and varied (Van der Krieken & De Graaf, 2015; Van Holsteyn & Vollaard, 2015). Municipalities and provinces can autonomously decide whether to adopt referendum legislation and have significant influence over the general rules of the game, including the initiator, signature requirements, number of options, voting method and use of quorums. Municipal referendum ordinances can specify that the municipal council can trigger referendums (the raadplegend referendum in Dutch), that a specific number of citizens’ signatures can qualify a proposal for a referendum (raadgevend referendum) or both (Hendriks et al., 2017). Around one third of Dutch municipalities have a referendum ordinance, as do around half of all provinces – although no provincial referendums have been held to date (Van der Krieken, 2015 – none were held since). On all government levels, a legislative majority can trigger an ad hoc referendum in the absence of referendum legislation.
  In Belgium, general regulations for referendums at local and provincial levels were introduced in 1995. Contrary to the Netherlands, the competence to regulate referendums is not decentralised to individual municipalities, and top-down as well as bottom-up triggered referendums, both reactive and proactive, are possible in all municipalities (Buelens, 2009). Of around 40 municipal referendums held by 2018, around half were triggered bottom-up – albeit predominantly in Flemish municipalities – and half were triggered top-town – albeit predominantly in Walloon municipalities (Goethals, 2018). Similarly to the Netherlands, no provincial referendums have been held under existing legislation (Bauwens, 2007). In updates to the legislation in 1999, signature requirements were diversified according to the number of inhabitants, de facto setting them higher. The 40% participation quorum was lowered to match the signature quorum – down to 10% in larger cities – and meeting the signature requirements now automatically qualifies a proposal for a referendum, as opposed to prior leeway for the municipal council to avert the vote. Regulations determine that referendums must involve a dichotomous (yes/no) question. Multiple questions can, however, be proposed. Municipal councils have the right to change referendum questions submitted by initiators or to add additional questions (Beckers & Billiet, 2009). In some instances this essentially produced political counter-proposals. The possibility for the political elites to determine the referendum question is subject to critique. Municipal councils sometimes alter the intent of the referendum initiators, such as in the 1999 local referendum in Gent, where a citizens’ initiative for free public transport was replaced by a ballot question on ‘better public transport’ (Bauwens, 2007). Besides not matching the intention of the initiators, the altered proposal also largely provided a quasi-choice to voters, as it would be hard to argue against better transportation. The same absence of two realistic competing alternatives applies to referendum votes on proposals that were unlikely to yield significant support, such as the construction of an incinerator in the municipality of Ciney in 1996 (Beckers, 2005).
  In Luxembourg, depending on the municipal size, 20–25 percent of the electorate can trigger a municipal referendum, but no bottom-up referendums have been initiated so far. Up until 2011, nine top-down referendums were triggered by municipal councils (Dumont et al., 2011). Seven of these referendums focused on municipal mergers, making this a popular topic for top-down referendums, similarly to the Netherlands.
  Considering the unlikely manifestation of national-level referendums in the Netherlands and Belgium in the near future, on the one hand, and the larger scope for innovation and experimentation at the local level, on the other, the remainder of this section will focus predominantly on existing and conceivable referendum innovations at the municipal level. Several suggestions made by advisory committees and individuals pioneering democratic innovations at national levels could nonetheless inspire the innovation of local referendum procedures in ways similar to those described in the previous section. The focus is on proposals that lead to a referendum vote, although creative proposals in the realms of delegating, amending or extending policy options in an informal voting process are also considered, as they could be adapted for application in a referendum process.

  5.1 Emerging and Potential Top-Down Referendum Innovations

  About 40% of municipal referendums in the Netherlands have included either more than two ballot options or multiple ballot questions (Van der Krieken, 2015). Most concerned votes on municipal mergers, with various merger scenarios presented to voters (Van der Krieken & De Graaf, 2015). Other examples include referendums on diverse harbour designs (in Arnhem in 2007), city centre redevelopment (in Duiven in 2008) and a new city hall (in Den Helder in 2013). In Flemish Belgium, just over half of all local referendums have entailed multiple ballot questions (Beckers & Billiet, 2009). In Wallonia, multiple scenarios were offered, for example, on art installations in Liège in 1995, redevelopment of the central square in Ath in 1995 and city development projects in Spa in 2007 (Beckers, 2005; Goethals, 2018).
  In the Netherlands, ballot options for top-down triggered referendums were either designed by politicians or delegated to experts or an external design panel. In some cases, societal inputs were included in the ballot design process. For example, in Arnhem, citizens could attend debates, participative events and target-group sessions to provide input on a longlist of possibilities. Based on the input, a shortlist was compiled for the referendum vote (Boogers & De Graaf, 2008). In an informal referendum in Borne, four visions for the future of the municipality, varying, among other things, in terms of population growth to be pursued, were presented to voters in a consultation with online, postal and physical voting opportunities. The four scenarios were derived from societal inputs through individual citizen interviews, focus groups and an online consultation. The inputs were filtered by a steering group in which 20 societal organisations were represented (Denters & Klok, 2015). Despite not involving an official referendum, the success of this experiment could inspire municipalities to design multi-option referendums based on citizen input. Consensual traditions in Dutch democracy provide fruitful experimentation opportunities for the incorporation of deliberative elements in referendum processes.
  Citizens’ assemblies (as discussed in Sections 4.1.1 and 4.1.2) could be instituted to prepare one or several scenarios or policy solutions for decision-making by the local electorate. Assemblies could be created ad hoc or follow a more institutionalised model. Useful lessons of such hybrid democratic innovations can be drawn from practices abroad, such as the Irish citizens’ assembly (Hendriks et al., 2020). The recent Dutch advisory state commission on citizen participation in climate policymaking concluded that, under the Irish model, the incorporation of citizen perspectives through deliberation and referendum voting culminated in policy innovation not deemed possible within the conventional decision-making process (Commissie Brenninkmeijer, 2021: 15). Inspiration can also be drawn from other innovations within the Netherlands and Belgium, which to date have not been linked to referendums. An example is the cooperative neighbourhood council in the Oosterpark neighbourhood in the city of Groningen (2017-2020), in which eleven randomly selected citizens and six local politicians were jointly responsible for decision-making in the neighbourhood. During each decision-making process, a panel of 400 randomly selected residents was consulted (Westerweel, 2021). A similar format could also involve referendum voting on preliminary ideas or final proposals. In the German-speaking part of Belgium, the Ostbelgien model employs a permanent council of randomly selected citizens, which are regularly rotated. The council is tasked with agenda-setting and delegates selected issues to topic-specific temporary citizens’ assemblies that formulate concrete policy recommendations. The recommendations are then discussed in a mixed committee of citizens, representatives and the responsible minister (Niessen & Reuchamps, 2022). This model could be elaborated to include the possibility to delegate the resulting proposal to a referendum, either directly (through top-down or even mandatory triggering) or indirectly, following the example of the German three-tier model.
  The application of multi-option or preferendum-style voting (see Sections 4.1.2 and 4.1.3) has also received attention in the Dutch and Belgian contexts.⁹x In societal discussions, the term ‘preferendum’ is often used for referendums entailing multiple options, not all of which necessarily entail preferential voting procedures and/or diverse ballot authors, as intended in the original definition of preferendums (as discussed in Section 4.1.3). Several of the suggestions discussed in this part thus bear more similarity to referendums with multiple ballot proposals as discussed in Section 4.1.2. The Commissie Brenninkmeijer (2021: 34) recommended the investigation of the possibility of a preferendum – to be formally be triggered by the parliament – on the proposals formulated by a citizens’ assembly, as offering multiple options would fit the plural nature of the assembly’s recommendations. An early-stage preferendum posing multiple policy scenarios on EU-related legislation had previously been advocated by the Netherlands Scientific Council for Government Policy (WRR, 2007) as a way to stimulate citizens to actively consider policy options and to result in more concrete policy consequences for policymakers. A proposal to legislate for a national-level preferendum triggered by parliamentarians was briefly on the agenda of ruling party VVD in the run-up to the most recent parliamentary elections but was revoked from the party manifesto by party members (NOS, 2020). The use of preferendums has more recently been promoted by Belgian cultural historian David Van Reybrouck. His conceptualisation of a preferendum takes a more elaborate form than voting on competing policy proposals contained within a single ballot question. He suggests a format in which citizens express their level of agreement and their top four priorities on a number of non-mutually exclusive proposals within the same topical realm (Van Reybrouck, 2021).A similar initiative is thematic voting, promoted by innovation researcher Rudy van Belkom, in which voters select their most preferred policies on a variety of policy domains from among five options each. In the spring of 2022, around the time of the municipal elections, the municipalities of The Hague and Breda experimented with thematic voting (De Jonge, 2021; Boonstra, 2022).While perhaps too elaborate to be translated into a formal referendum ballot, these initiatives can provide inspiration for either extending ballot choice beyond a single proposal in top-down referendums or for more informal applications of multi-option balloting as a tool for citizen participation and input. The essential question remains who designs the alternatives for the ballot.

  5.2 Emerging and Potential Bottom-Up Referendum Innovations

  The advisory state commission on the parliamentary system recommended allowing reactive referendums to target separate elements of legislation, as opposed to legislation in full (Staatscommissie parlementair stelsel, 2018: 147). Bottom-up referendums in Dutch municipalities, both proactive and reactive, currently tend to have a binary nature. Replacing binary corrective referendums with semi-propositive referendums (see Section 4.2.2.1) could prevent legislation being abolished over dissent with specific elements. Randomly selected citizens’ assemblies could also play a role in articulating the discordances with public opinion that ought to be addressed (Hendriks, 2021) or in formulating a citizens’ counter-proposal (see Section 4.2.2.2).
  The city of Amsterdam renewed its referendum ordinance in 2022 to include amendments and counter-proposals. Similarly to the German three-tier model (see Section 4.2.1.1), a citizens’ initiative with the initial support of 1,000 citizen signatures can be adopted by the municipal council, in which case no referendum is held. The council can also propose amendments to the initiative, which are discussed with the initiators. If accepted, the amended initiative is turned into legislation. If the initiators do not accept the amendments, 10,000 citizen signatures qualify the initiative for a referendum vote, to which the municipal council can propose a political counter-proposal (see Section 4.2.1.2). For reactive referendums, initiators can collect signatures either to abolish a legislative proposal or to support their citizens’ counter-proposal (see Section 4.2.2.2). In the latter case, the same three outcomes are possible: adoption by the council, a compromise between council and initiators, or the collection of 10,000 signatures for a referendum vote.¹⁰x Referendum ordinance city of Amsterdam 2022: https://lokaleregelgeving.overheid.nl/CVDR671705/1. The first city-wide referendum under this new legislation is yet to be held, but the legal provision for counter-proposals can be considered innovative in the Dutch context. Amsterdam has some prior experience with a counter-proposal referendum, although it was not legally regulated as such. A 2014 district referendum in central Amsterdam posed a municipal proposal for redevelopment of a street against a rival proposal formulated by residents. The referendum was organised after 4,000 residents petitioned for a citizens’ initiative under district-level referendum legislation valid at the time.¹¹x Referendum ordinance district of central Amsterdam (2013-2015): https://lokaleregelgeving.overheid.nl/CVDR74030/3. The council decided to pose both proposals against each other in a single referendum. The municipal proposal received a 55% majority of votes. The relatively narrow margin in votes indicates the relevance of including two competing options.
  Existing critiques on Belgian bottom-up referendum procedures (e.g. Bauwens, 2007; Beckers, 2005), in which the referendum question can be altered by the municipal council, could be addressed by legislating that changes to the question must either be accepted by the initiators – similarly to the three-tier model (see Section 4.2.1.1) – or take the form of a municipal counter-proposal (see Section 4.2.1.2), thus presenting an additional ballot option. Already in 2009, Belgian advocacy group Meer Democratie suggested the introduction of a three-tier model in Belgium.¹²x Press release on their proposal: www.meerdemocratie.be/persbericht-van-democratienu. Another benefit of this model would be the possibility for the municipal council to adopt the initiative without a referendum vote. This could have avoided referendums in several Flemish municipalities where the municipal majority supported and actively campaigned for the policy proposal (Beckers & Billiet, 2009). Improving the quality of ballot questions could also help to ensure a realistic choice between different alternatives and to prevent distortion of the referendum question by political elites. Beckers (2005; see also Beckers & Billiet, 2009) recommends the implementation of a multi-step procedure in which referendum initiators work towards a suitable referendum question in consultation with the advisory committee for referendums (the Vlaamse Adviescommissie voor Volksraadplegingen, VAV).¹³x This Flemish body advises, on request, on the referendum question for local referendums and on various organisational aspects (Beckers & Billiet, 2009). Its role bears similarity to Dutch referendum commissions, except that the latter are decentralised at the local level (Van Praag, 2009). Such commissions of external, independent referendum experts are hailed by the Association of Netherlands Municipalities (VNG) as an important contribution to decent referendum processes: https://vng.nl/sites/default/files/2019-12/20191210_ledenbrief_nieuwe-vng-model-referendumverordening.pdf. In both Belgium and the Netherlands, existing advisory committees of independent experts could fulfil a supporting role in optimising the proposal presented on the referendum ballot in the absence of judicial review procedures common to the US context (see Section 4.2.1.3). Involving independent experts could curtail the issue of double agendas for the municipal council as both legislator and responsible body for referendum organisation (Van Praag, 2009).
  The advisory state commission on the parliamentary system suggested the linking of a citizens’ assembly to new requests for what is called the popular initiative at the Dutch national level.¹⁴x Contrary to citizens’ initiatives as defined in Section 2, the national-level Dutch popular initiative is an agenda-setting instrument, in which a minimum number of 40,000 citizens can request parliament to discuss a proposal. See parliamentary website: www.tweedekamer.nl/kamerleden_en_commissies/commissies/verz/burgerinitiatieven. The assembly would develop policy options in consultation with society (Staatscommissie parlementair stelsel, 2018: 167-168). Although the commission advises against following up with a referendum at the national level considering the lack of experience, this procedure could be worthy of experimentation at the local level as a variation on the citizens’ assembly situated between drafting and signature collection (see Section 4.2.1.5). Also providing inspiration is a parliamentary initiative to introduce a citizens’ amendment procedure.¹⁵x Parliamentary website: www.tweedekamer.nl/kamerstukken/detail?id=2019Z22328&did=2019D46412. With 70,000 signatures, citizens could propose changes to legislation currently being discussed in parliament. Such a procedure would bear similarity to a citizens’ counter-proposal or semi-propositive referendum (see Sections 4.2.2.1 and 4.2.2.2) if it were combined with a referendum vote.
  In 2016, a newspaper competition for democratic innovations featured Jan Veneman’s proposal for ‘Voters’ choice’, with citizen signatures initiating a procedure in which the electorate votes on a policy domain in between parliamentary elections. Experts and stakeholders would design three policy scenarios for this domain. The most supported scenario would form the basis for parliamentary policymaking. This process – similar to thematic voting but triggered bottom-up and on a single policy domain – would facilitate the continuous adjustment or redirection of policies (Veneman, 2016). Veneman’s proposal does not directly fit any of the models presented in Section 4, being triggered bottom-up yet bypassing initiators in formulating ballot options, though it does contain elements of various suggested innovations (for example those discussed in Sections 4.1.2, 4.1.3 and 4.2.2.1). Alternatively, the procedure could be adapted to include various citizen-designed proposals (see Section 4.2.1.4). Having been dismissed as difficult to implement at the national level, the procedure can provide inspiration for local level innovations, where such initiatives are more practical to organise.
  In conclusion, the bottom-up referendum is a commonly practised referendum type in Dutch and Belgian municipalities that, despite its general popularity, can pose challenges to local decision-making, especially when referendum questions do not pose a genuine choice to voters or when uncertainty prevails over whether and how to adapt rejected legislation. Partial rejection, citizens’ amendments and counter-proposals could make the corrective referendum process more constructive and enhance dialogue between initiators, legislators and the wider electorate, particularly when a randomly selected citizens’ assembly is embedded in the referendum process. Likewise, by incorporating stages of political or societal feedback and room for amendments in the process, proactive citizens’ initiatives could build on broader support than the interests of their initiators. The low countries have no tradition with judicial review of initiatives, but municipal referendum commissions or regional referendum authorities like the VAV could take up a reviewing role of initiatives, amendments and the incorporation of societal feedback.

• 6 Conclusion

  This article reviewed innovations to referendum processes, in which delegating authorship, integrating feedback and amendments and admitting additional ballot options are expected to improve ballot choice in line with societal preferences. By innovating referendums in such a way that ballot proposals are not immutable after triggering but can be amended, or supplemented with additional alternatives, citizens can exert influence over not just the topic but also the direction of new policy. This reduces the power of initiators to formulate referendum proposals in line with their own interests and decreases instances of legislation being rejected over amendable objections. Local and international experiences as well as theoretical and practical suggestions by academics, advisory committees, parliamentarians and individuals, as discussed in this article, could all provide inspiration for experimentation with innovative referendum procedures in the low countries and beyond.
  Implementing procedural innovations in referendums would benefit from four general recommendations. The first is to take into account the context, such as prior experience with referendums, the institutionalisation and prevalence of societal deliberation and provisions for judicial or expert review. The second recommendation is to properly regulate referendum procedures. For more innovative referendum procedures, particularly when bottom-up triggered, legislation specifying the conditions, actors and procedural steps is essential and must be communicated to actors within the legislature and society. Third, new procedures for referendums require new skills from those involved and take time to become embedded in the democratic system. By experimenting with innovations, they can be improved to include unforeseen procedural and contextual details. The fourth recommendation is to evaluate experiences and satisfaction among citizens and legislators in order to further improve the process. Experiences can be shared with other legislative entities for inspiration. Academic research also plays a role in disseminating innovative procedures and their evaluations, comparing different models and building a broader knowledge base of what works in different situations or contexts.
  Bearing in mind these recommendations, the referendum innovations discussed in this article could provide new impulses to citizen participation and referendum voting in the low countries and elsewhere. They could be supplemented by other innovative instruments, for example in the realms of societal deliberation, voter education and outcome monitoring of referendum results. Allowing amendments to, and extensions of, ballot content could be fruitful steps in making referendums more constructive instruments and engaging larger segments of society in the policymaking process.

References

• Akkerman, T. (2004). Talking, Voting and Violence. Divisive Issues and the Limits of Democratic Deliberation. Amsterdam School for Social Science Research working paper 04/05.

• Altman, D. (2011). Direct Democracy Worldwide. Cambridge: Cambridge University Press.

• Altman, D. (2014). Strengthening Democratic Quality: Reactive Deliberation in the Context of Direct Democracy. Working Paper Kellogg Institute for International Studies, 400.

• Altman, D. (2018). Citizenship and Contemporary Direct Democracy. Cambridge University Press.

• Barber, B. (1984). Strong Democracy: Participatory Politics for a New Age. University of California Press.

• Bauwens, M. (2007). België en de directe democratie. In J. Verhulst & A. Nijeboer (Eds.), Directe democratie Feiten, argumenten en ervaringen omtrent het referendum (pp. 92-97). Democracy International.

• Beckers, M. (2005). De vraagstelling in volksraadplegingen en referenda. Literatuurstudie en analyse van het maatschappelijk debat a.d.h.v. een aantal cases. Thesis KU Leuven.

• Beckers, M. & Billiet, J. (2009). ‘Handhaven’ of ‘herroepen’? De vraagstelling in twaalf gemeentelijke volksraadsplegningen in Vlaanderen onderzocht. Res Publica, 51(1), 53-82.

• Bjørklund, T. (1982). The Demand for Referendum: When Does It Arise and When Does It Succeed? Scandinavian Political Studies, 5, 237-259.

• Bochsler, D. (2010). The marquis de Condorcet goes to Bern. Public Choice, 144(1-2), 119-131.

• Boogers, M. J. G. J. A. & de Graaf, L. (2008). Een ongewenst preferendum. Tilburg: Tilburgse School voor Politiek en Bestuur.

• Boonstra, W. (2022). Veel animo voor themastemmen bij digitale schaduwverkiezing. Binnenlands Bestuur, 10 May. www.binnenlandsbestuur.nl/bestuur-en-organisatie/bredase-themastemmers-kiezen-socialer.

• Breuer, A. (2008). Policymaking by Referendum in Presidential Systems: Evidence from the Bolivian and Colombian Cases. Latin American Politics and Society, 50(4), 59-89.

• Broder, D. S. (2000). Democracy Derailed: Initiative Campaigns and the Power of Money. Houghton Mifflin.

• Buelens, J. (2009). Volksraadsplegingen: kan België wat leren van Nederland? Res Publica, 51(1), 13-32.

• Caluwaerts, D. & Reuchamps, M. (2020). Still Consociational? Belgian Democracy, 50 Years After ‘The Politics of Accommodation’. Politics of the Low Countries, 2(1), 54-76.

• Capretti, A. (2002). Reformperspektiven zum abrogativen Referendum in Italien. In T. Schiller & V. Mittendorf (Eds.), Direkte Demokratie: Forschung und Perspektiven (pp. 305-315). Westdeutscher Verlag.

• Commissie Brenninkmeijer (2021). Betrokken bij klimaat. Burgerfora aanbevolen. Eindrapportage adviescommissie Burgerbetrokkenheid bij klimaatbeleid, Rijksoverheid. Retrieved from https://www.rijksoverheid.nl/documenten/publicaties/2021/03/21/adviesrapport-betrokken-bij-klimaat, last accessed 7 February 2023.

• De Jonge, A. (2021). Den Haag gaat pionieren met themastemmen. Binnenlands Bestuur, 22 December. www.binnenlandsbestuur.nl/bestuur-en-organisatie/den-haag-wil-experiment-met-thematisch-stemmen

• Denters, S. A. & Klok, P. J. (2015). MijnBorne2030: evaluatie van een democratisch experiment. Bestuurswetenschappen, 69(2), 24-45. doi: 10.5553/Bw/016571942015069002004.

• Dumont, P., Kies, R. & Poirier, P. (2011). Luxembourg: The Challenge of Inclusive Democracy in a ‘Local State’. In J. Loughlin, F. Hendriks & A. Lidström (Eds.), The Oxford Handbook of Local and Regional Democracy in Europe. Oxford University Press.

• Emerson, P. (1993). Letter to the Editor. Representation, 31(116), 87. doi: 10.1080/00344899338439021.

• Emerson, P. (2021). Democratic Decision-Making: Consensus Voting for Civic Society and Parliaments. Springer International Publishing.

• Farrell, D. M., Suiter, J. & Harris, C. (2019). ‘Systematizing’ Constitutional Deliberation: The 2016-18 Citizens’ Assembly in Ireland. Irish Political Studies, 34(1), 113-123. doi: 10.1080/07907184.2018.1534832.

• Ferejohn, J. (2008). The Citizens’ Assembly Model. Designing Deliberative Democracy: The British Columbia Citizens’ Assembly, 192-213. doi: 10.1017/CBO9780511491177.011.

• Gallagher, J. (2014). Scotland’s Future – Really? Renewal: A Journal of Social Democracy, 22(1/2), 139-145.

• Gallagher, M. & Uleri, P. V. (Eds.). (1996). The Referendum Experience in Europe. Macmillan.

• Gastil, J. & Richards, R. (2013). Making Direct Democracy Deliberative Through Random Assemblies. Politics & Society, 41(2), 253-281. doi: 10.1177/0032329213483109.

• Geissel, B. (2016). Direct Democracy and its (Perceived) Consequences: The German Case. In S. Ruth-Lovell, Y. Welp & L. Whitehead (Eds.), Let the People Rule: Direct Democracy in the Twenty-First Century. ECPR Press.

• Glaser, A., Serdült, U. & Somer, E. (2016). Das konstruktive Referendum–ein Volksrecht vor dem Aus? Aktuelle Juristische Praxis (AJP), 10, 1343-1355.

• Goethals, A. (2018). Een rechtsvergelijkend onderzoek naar de lokale volksraadpleging in België en het lokale referendum in Nederland. Master’s thesis Universiteit Gent.

• Hendriks, F. (2021). Het ingelote burgerforum: observaties uit het internationale debat en handelingsperspectieven voor Nederland. In F. Hendriks, K. Jacobs & A. Michels (Eds.), Nationale Burgerfora. Verkenning van nationale burgerfora als democratisch gereedschap (pp. 6-23). Ministerie van Binnenlandse Zaken en Koninkrijksrelaties.

• Hendriks, F., Jacobs, D. J. & Wagenaar, C. C. L. (2020). Referenderen en delibereren. Een verkenning van bindend correctieve referenda en combinaties met deliberatieve democratie. Tilburg University.

• Hendriks, F., Van der Krieken, K. H. J. & Wagenaar, C. C. L. (2017). Democratische zegen of vloek? Aantekeningen bij het referendum. Amsterdam University Press.

• Hug, S. & Tsebelis, G. (2002). Veto Players and Referendums Around the World. Journal of Theoretical Politics, 14(4), 465-515. doi: 10.1177/095169280201400404.

• Jacobs, K. (Ed.). (2018). Het Wiv-referendum: Nationaal Referendum Onderzoek 2018. Stichting KiezersOnderzoek Nederland (SKON).

• Jäske, M. (2017). “Soft” Forms of Direct Democracy. Explaining the Occurrence of Referendum Motions and Advisory Referendums in Finnish Local Government. Swiss Political Science Review, 1, 50-76. doi: 10.1111/spsr.12238.

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

• Kampwirth, R. (2004). Direct Democracy in Germany. In B. Kaufmann & M. D. Waters (Red.), Direct Democracy in Europe: A Comprehensive Reference Guide to the Initiative and Referendum Process in Europe. Carolina Academic Press.

• Kaufmann, B. & Waters, D. (Eds.). (2004). Direct Democracy in Europe: A Comprehensive Reference Guide to the Initiative and Referendum Process in Europe. Carolina Academic Press.

• Lagerspetz, E. (2016). Social Choice and Democratic Values. Springer.

• Landemore, H. (2015). Inclusive Constitution-Making: The Icelandic Experiment. The Journal of Political Philosophy, 23(2), 166-191. doi: 10.1111/jopp.12032.

• Levine, S. & Roberts, N. S. (1993). The New Zealand Electoral Referendum of 1992. Electoral Studies, 12(2), 158-167. doi: 10.1016/0261-3794(93)90015-C.

• Lundberg, T. C. (2007). Electoral System Reviews in New Zealand, Britain and Canada: A Critical Comparison. Government and Opposition, 42(4), 471-490. doi: 10.1111/j.1477-7053.2007.00232.x.

• Lupia, A. & Johnston, R. (2001). Are Voters to Blame? Voter Competence and Elite Maneuvers in Referendums. In M. Mendelsohn & A. Parkin (Eds.), Referendum Democracy (pp. 191-210). Palgrave Macmillan.

• Mac Ginty, R. (2003). Constitutional Referendums and Ethnonational Conflict: The Case of Northern Ireland. Nationalism and Ethnic Politics, 9(2), 1-22. doi: 10.1080/13537110412331301395.

• Mackerras, M. (1994). Reform of New Zealand’s Voting System, 1985-1996. Representation, 32(118), 36-40. doi: 10.1080/00344899438439041.

• Magleby, D. B. (1995). Let the Voters Decide? An Assessment of the Initiative and Referendum Process. University of Colorado Law Review, 66, 13.

• McCarty Carino, M. (2018). How California Initiatives Went from ‘Power To The People’ To A Big Money Game. Capital Public Radio, 1 November. www.capradio.org/articles/2018/11/01/how-california-initiatives-went-from-power-to-the-people-to-a-big-money-game/.

• McKay, S. (2018). Designing Popular Vote Processes for Democratic Systems: Counter-Proposals, Recurring Referendums, and Iterated Popular Votes. Swiss Political Science Review, 24(3), 328-334. doi: 10.1111/spsr.12316.

• McLean, I., Spirling, A. & Russell, M. (2003). None of the Above: The UK House of Commons Votes on Reforming the House of Lords, February 2003. The Political Quarterly, 74(3), 298-310. doi: 10.1111/1467-923X.00539.

• Mendelsohn, M. & Parkin, A. (2001). Introduction: Referendum Democracy. In M. Mendelsohn & A. Parkin (Eds.), Referendum Democracy (pp. 1-22). Palgrave Macmillan.

• Michels, A. (2011). Innovations in Democratic Governance: How Does Citizen Participation Contribute to a Better Democracy? International Review of Administrative Sciences, 77(2), 275-293. doi: 10.1177/0020852311399851.

• Mitchell, J. (1992). The Multi-option Referendum: A Comparative Perspective. In A. Macartney (Ed.), Asking the People: The Referendum and Constitutional Change (pp. 7-19). Unit for the Study of Government in Scotland, Edinburgh University.

• Morel, L. (2018a). The Democratic Criticism of Referendums. In L. Morel & M. Qvortrup (Eds.), The Routledge Handbook to Referendums and Direct Democracy (pp. 149-168). Routledge.

• Morel, L. (2018b). Types of Referendums, Provisions and practice at the National Level Worldwide. In L. Morel & M. Qvortrup (Eds.), The Routledge Handbook to Referendums and Direct Democracy (pp. 27-59). Routledge.

• Morison, J. & Newman, D. R. (2001). On-line Citizenship: Consultation and Participation in New Labour’s Britain and Beyond. International Review of Law, Computers & Technology, 15(2), 171-194. doi: 10.1080/13600860120070501.

• Nagel, J. H. (1994). What Political Scientists Can Learn from the 1993 Electoral Reform in New Zealand. PS:Political Science and Politics, 27(3), 525-529. doi: 10.2307/420221.

• Niessen, C. & Reuchamps, M. (2022). Institutionalising Citizen Deliberation in Parliament: The Permanent Citizens’ Dialogue in the German-speaking Community of Belgium. Parliamentary Affairs, 75(1), 135-153. doi: 10.1093/pa/gsaa056.

• Nikolenyi, C. (2011). When Electoral Reform Fails: The Stability of Proportional Representation in Post-Communist Democracies. West European Politics, 34(3), 607-625. doi: 10.1080/01402382.2011.555986.

• NOS (2020). VVD-leden: praten met GroenLinks mag, ‘preferendum’ sneuvelt. 12 December. https://nos.nl/artikel/2360325-vvd-leden-praten-met-groenlinks-mag-preferendum-sneuvelt.

• O’Flynn, I. & Levy, R. (2020). Deliberative Constitutional Referendums in Deeply Divided Societies. University of British Columbia Law Review, 53(1), 205-240.

• Orr, G. (2001). Preferenda: The Constitutionality of Multiple Option Referenda. Constitutional Law and Policy Review, 3(4), 68-72.

• Papadopoulos, Y. (1995). Analysis of Functions and Dysfunctions of Direct Democracy: Top-Down and Bottom-Up Perspectives. Politics & Society, 23(4), 421-448. doi: 10.1177/0032329295023004002.

• Parkinson, J. (2001). Deliberative Democracy and Referendums. In K. Dowding, J. Hughes & H. Margetts (Eds.), Challenges to Democracy (pp. 131-152). Palgrave Macmillan.

• Rein, M. (2008). Reframing Problematic Policies. In M. Moran, M. Rein & R. E. Goodin (Eds.), The Oxford Handbook of Public Policy (pp. 389-405). Oxford University Press.

• Rohner, A. (2011). Direct Democracy in the German Länder: History, Institutions and (Mal) Functions. C2D Working Paper series.

• Ruin, O. (1982). Sweden in the 1970s. In J. Richardson (Ed.), Policy Styles in Western Europe (pp. 141-168). George Allen & Unwin.

• Scarrow, S. E. (2001). Direct Democracy and Institutional Change: A Comparative Investigation. Comparative Political Studies, 34(6), 651-665. doi: 10.1177/0010414001034006003.

• Schacter, J. S. (1995). The Pursuit of “Popular Intent”: Interpretive Dilemmas in Direct Democracy. The Yale Law Journal, 105(1), 107-176. doi: 10.2307/797141.

• Şen, I. G. (2015). Sovereignty Referendums in International and Constitutional Law. Springer.

• Setälä, M. (1999). Referendums and Democratic Government: Normative Theory and the Analysis of Institutions. Macmillan Press.

• Smith, G. C. (2007). Solving the Initiatory Construction Puzzle (and Improving Direct Democracy) by Appropriate Refocusing on Sponsor Intent. University of Colorado Law Review, 78, 257-305.

• Staatscommissie parlementair stelsel (2018) Lage drempels, hoge dijken: democratie en rechtsstaat in balans. Boom.

• Staszewski, G. (2003). Rejecting the Myth of Popular Sovereignty and Applying an Agency Model to Direct Democracy. Vanderbilt Law Review, 56(2), 395-496. doi: 10.2139/ssrn.405520.

• Stefanini, M. F. R. (2018). Referendums, Minorities and Individual Freedoms. In L. Morel & M. Qvortrup (Eds.), The Routledge Handbook to Referendums and Direct Democracy (pp. 371-387). Routledge.

• Suksi, M. (1993). Bringing in the People: A Comparison of Constitutional Forms and Practices of the Referendum. Martinus Nijhoff Publishers.

• Taillon, P. (2018). The Democratic Potential of Referendums. In L. Morel & M. Qvortrup (Eds.), The Routledge Handbook to Referendums and Direct Democracy (pp. 169-191). Routledge.

• Tierney, S. (2013). The Multi-Option Referendum: International Guidelines, International Practice and Practical Issues. Edinburgh: University of Edinburgh, School of Law, Working Papers.

• Tsebelis, G. (2018). How Can We Keep Direct Democracy and Avoid “Kolotoumba”. Homo Oeconomicus, 35(1-2), 81-90. doi: 10.1007/s41412-018-0069-4.

• Tudor, K. (2016). Raising (Issues About) the Banner: A Critical Reflection on the New Zealand Flag Debate. Counterfutures, 2, 83-112. doi: 10.26686/cf.v2i0.6431.

• Uleri, P. V. (1996). Introduction. In M. Gallagher & P.V. Uleri (Eds.). The referendum experience in Europe (pp. 1-19). Macmillan Press.

• Uleri, P. V. (2002). On Referendum Voting in Italy: YES, NO or Non–vote? How Italian Parties Learned to Control Referendums. European Journal of Political Research, 41(6), 863-883. doi: 10.1111/1475-6765.t01-1-00036.

• Van der Krieken, K. H. J. (2015). Het lokale referendum in Nederland: Een verkenning van de lokale referendumpraktijk in Nederland en scenario’s voor de toekomst. Tilburg University.

• Van der Krieken, K. H. J. & De Graaf, L. (2015). Dichotoom in dialoog? Lokale referenda in Nederland, Bestuurswetenschappen, 69(2). doi: 10.5553/Bw/016571942015069002005.

• Van der Meer, T. W. G., Wagenaar, C. C. L. & Jacobs, K. T. E. (2020). The Rise and Fall of the Dutch Referendum Law (2015-2018): Initiation, Use, and Abolition of the Corrective, Citizen-Initiated and Non-binding Referendum. Acta Politica. Advance online publication. doi: 10.1057/s41269-020-00175-3.

• Van Holsteyn, J. & Vollaard, H. (2015). Ratifying the European Constitutional Treaty by Referendum. The End of Consensus Politics? In H. Vollaard, J. Beyers & P. Dumont (Eds.), European Integration and Consensus Politics in the Low Countries (pp. 213-233). Routledge.

• Van Praag, P. (2009). Onafhankelijke referendumcommissies: kenmerkend voor de Nederlandse consensusdemocratie. Res Publica, 51(1), 33-52. doi: 10.5553/RP/048647002009051001003.

• Van Reybrouck, D. (2021). Geef burgers échte invloed op klimaatbeleid. NRC Handelsblad, 8 October.

• Veneman, J. (2016). Geef kiezers echt wat te kiezen. NRC Handelsblad, 22 October.

• Venice Commission. (2007). Code of Good Practice on Referendums. CDL-AD(2007)008rev-cor.

• Verhulst, J. & Nijeboer, A. (2007). Direct Democracy: Facts and Arguments about the Introduction of Initiative and Referendum. Democracy International.

• Wagenaar, C. C. L. (2019). Beyond for or Against? Multi-option Alternatives to a Corrective Referendum. Electoral Studies, 62, 102091. doi: 10.1016/j.electstud.2019.102091.

• Wagenaar, C. C. L. (2020). Lessons from International Multi-option Referendum Experiences. The Political Quarterly, 91(1), 192-202. doi: 10.1111/1467-923X.12823.

• Wagenaar, C. C. L. (2021). Voting beyond Vetoing: Variations in Agenda-Setting and Balloting Procedures for Multi-option Referendums. PhD thesis, Tilburg University.

• Wagenaar, C. C. L. & Hendriks, F. (2021). Setting the Voting Agenda for Multi-option referendums: Process Variations and Civic Empowerment. Democratization, 28(2), 372-393. doi: 10.1080/13510347.2020.1823969.

• Warren, M. E. & Pearse, H. (2008). Designing Deliberative Democracy: The British Columbia Citizens’ Assembly. Cambridge University Press.

• Westerweel, J. (2021). Lokale democratische innovatie. Een juridische analyse. PhD thesis, Leiden University.

• Wetenschappelijke Raad voor het Regeringsbeleid (WRR). (2007). Europa in Nederland. Amsterdam University Press.

Noten

• 1 Facultative referendums contrast with mandatory referendums. In some countries, such as Ireland, the constitution stipulates that the electorate must approve any changes to the constitution in a referendum. The referendum is thus said to be triggered automatically, contrary to the facultative, i.e. voluntary, triggering of a referendum by particular actors (e.g. Breuer, 2008). Since mandatory triggering does not actively involve referendum initiators, such referendums fall outside the scope of the theoretical framework for this article.

• 2 Others have used different terminologies for proactive or reactive characteristics, such as propositive versus non-propositive (Morel 2018b) or decision-promoting versus decision-controlling (Uleri, 1996).

• 3 For example, in various Latin American countries, both the president and parliament have legislative as well as referendum-triggering powers, which can result in a top-down triggered, reactive referendum on a proposal formulated by the other entity (Breuer, 2008). These less common referendum types are not the main focus of this article but could arguably be subject to similar innovations as those discussed for bottom-up triggered reactive referendums.

• 4 In the case of Ireland, referendums on constitutional changes are mandatory. In order to revoke the prohibition of abortion from the constitution and to allow parliament to legislate on abortion, popular approval in a referendum was obligatory. Since the citizens’ assembly could, however, not formally pass legislation, the parliament acted as an intermediator by formally proposing to change the constitution and to legislate for abortion in line with the assembly’s proposal.

• 5 This step is skipped under the two-tier process. Initiators can proceed to the second stage, collecting the additional signatures required to qualify the proposal for the ballot as soon as the initial request is declared admissible. This procedure is similar to bottom-up referendum procedures in, for example, the Netherlands, where the signature collection process is split into two phases, allowing for a determination of the admissibility of the referendum proposal before all signatures are collected. Under such a model, there is no possibility for the legislative to adopt the initiative or suggest amendments.

• 6 This distinguishes the German three-tier model from so-called ‘referendum motions’, which are used, for instance, at local levels in Finland, Norway and Iceland. Referendum motions obligate a municipal council to consider organising a referendum on a policy topic if proposed by a specified number of citizens, but initiators cannot force the vote if legislators decide against it. They are also referred to as ‘soft’ forms of direct democracy (Jäske, 2017).

• 7 The staging would entail a first round of voting on whether to retain or change the status quo, in case of majority support for change, a second stage in which the alternative proposals compete for support.

• 8 Parliamentary vote: www.tweedekamer.nl/kamerstukken/stemmingsuitslagen/detail?id=2022P13194&did=2022P13194.

• 9 In societal discussions, the term ‘preferendum’ is often used for referendums entailing multiple options, not all of which necessarily entail preferential voting procedures and/or diverse ballot authors, as intended in the original definition of preferendums (as discussed in Section 4.1.3). Several of the suggestions discussed in this part thus bear more similarity to referendums with multiple ballot proposals as discussed in Section 4.1.2.

• 10 Referendum ordinance city of Amsterdam 2022: https://lokaleregelgeving.overheid.nl/CVDR671705/1.

• 11 Referendum ordinance district of central Amsterdam (2013-2015): https://lokaleregelgeving.overheid.nl/CVDR74030/3.

• 12 Press release on their proposal: www.meerdemocratie.be/persbericht-van-democratienu.

• 13 This Flemish body advises, on request, on the referendum question for local referendums and on various organisational aspects (Beckers & Billiet, 2009). Its role bears similarity to Dutch referendum commissions, except that the latter are decentralised at the local level (Van Praag, 2009). Such commissions of external, independent referendum experts are hailed by the Association of Netherlands Municipalities (VNG) as an important contribution to decent referendum processes: https://vng.nl/sites/default/files/2019-12/20191210_ledenbrief_nieuwe-vng-model-referendumverordening.pdf.

• 14 Contrary to citizens’ initiatives as defined in Section 2, the national-level Dutch popular initiative is an agenda-setting instrument, in which a minimum number of 40,000 citizens can request parliament to discuss a proposal. See parliamentary website: www.tweedekamer.nl/kamerleden_en_commissies/commissies/verz/burgerinitiatieven.

• 15 Parliamentary website: www.tweedekamer.nl/kamerstukken/detail?id=2019Z22328&did=2019D46412.