Referendums, Initiatives, and Voters’ Accountability

Do democratic systems that include binding referendum and initiative processes violate a core principle of democracy, namely that legislators should be accountable? Some have argued that they do: these popular vote processes would grant the right to legislate to ordinary voters even though they cannot be held accountable—i.e., face possible consequences imposed by others for their actions and decisions. As a result, we should favor conventional representative systems over systems with popular vote processes. In this article, I analyze this ‘accountability objection’ to referendums and initiatives and the conception of voters as ‘legislators’ on which it relies and argue that it does not withstand scrutiny. I first demonstrate that there is no sound basis to characterize only voters in popular votes, and not voters in elections, as ‘legislators’. Therefore, the scope of the objection is broader than anticipated by its proponents: it cannot be consistently used to dismiss popular votes without also being fatal to elections. I then argue that voters’ lack of accountability is not fatal to either mass voting process. The institutional role of ordinary voters provides them with co-lawmaking powers; but these are not the kind of special powers that justify holding certain actors accountable in democratic systems. This does not exclude that voters have responsibilities. I close by highlighting that, compared to conventional representative systems, democratic systems including well-designed popular vote processes could offer voters more supportive conditions to meet such possible responsibilities.


Introduction: The Accountability Objection to Popular Vote Processes
Democratic theorists have proposed many reasons why popular vote processesall types of referendum and initiative processes that involve mass votes on political issues-should not be included in the range of democratic decision-making methods. 1 In this article, I focus on one of these reasons, which provides an opportunity to gain a better understanding of the role of ordinary voters in popular votes and of mass votes in democratic systems more generally.I will call it the accountability objection to popular vote processes.Often assumed, but rarely discussed (exceptions include Hamilton 1997;Trechsel 2010;Landa and Pevnick 2020), it can be reconstructed as follows.
In democratic systems, legislators should be accountable.This principle is implemented in conventional representative systems, in which legislators are elected and face public scrutiny and elections. 2In contrast, it is not implemented in systems that also include (any kind of) binding referendum and/or initiative processes.This is because the latter systems formally grant the right to legislate not only to elected actors, who are accountable, but also to ordinary voters, who are not.In the words of proponents of the objection: these processes would allow 'individual voters to make the law directly from the voting booth' (Hamilton 1997, p. 2) but provide no way to impose consequences onto them if they make bad or irresponsible decisions.Voters 'in a direct democracy […] are not stripped of valuable privileges for failing to use their power in a way that citizens think advances their interests' or for voting in uninformed or prejudiced ways (Landa and Pevnick 2020, p. 4)-and no one can 'dissolve the people and elect another' after a popular vote decision on a policy issue has been made, to cite Brecht's poem (Brecht's 1998, p. 440).The conception of accountability implied here is thus the possibility to face consequences imposed by others for one's actions and decisions-consequences such as answerability, the possibility to be called to publicly justify one's actions and decisions, or other forms of sanctions such as not being re-elected, seeing one's financial or organizational resources withheld, having one's decisions blocked or amended, and losing status, support, reputation, and relationships (see, e.g., Bovens 2007, pp. 451-452;Warren 2014, pp. 39-41). 3In the absence of possible consequences, due in great part to the secret ballot (Serota andLeib 2013, p. 1615), voters in a popular vote will use their legislative power badly.They 'can lie about [their] vote with impunity', have no incentives to get informed about the issues at hand, and do not need to articulate reasons for their decisions: for Hamilton, 'if there were ever a scenario when prejudice and foolishness could triumph, this would be it' (Hamilton 1997, p. 12).Landa and Pevnick, insist that because they are not accountable, 'ordinary citizens in a direct 1 See el-Wakil and McKay (2020) for a note on this terminology. 2I use the term 'conventional representative system' to refer to the set of institutions of electoral representative democracy without popular vote processes (which can take many different forms), not to an ideal of representative democracy.Furthermore, I assume that these systems, just as systems with popular vote processes, qualify as 'democratic' in the sense that they implement the three normative principles of empowered inclusion, collective agenda-setting and will formation, and collective decision-making to a sufficient extent (following Warren 2017). 3The accountability objection would make no sense if accountability was understood as accountability to one's self or one's own beliefs, as spontaneous mutual accountability, or as accountability in terms of bearing the consequences of one's choices (see, e.g., Mansbridge 2003, p. 520;Cheneval 2015, p. 167;Lafont 2020, p. 209), since such forms of voters' accountability that do not require the involvement of others are readily available.Such 'broad' conceptions of accountability (Bovens 2007, p. 450) are thus outside of considerations here.democracy' lack incentives 'to exercise their power with a level of care and consideration' as well as responsibly (Landa and Pevnick 2020, p. 5).The objection concludes that systems including binding popular vote processes should be rejected and that democracy should instead be implemented through conventional representative systems (Hamilton 1997; Landa and Pevnick 2020, pp.4-5). 4 In this article, I take on board the premise that ordinary voters cannot be held accountable (in the sense proposed by the objection), but I reject the conclusion that this justifies dismissing binding referendums and initiatives and implementing democracy through conventional representative institutions. 5My starting point is a perplexing aspect of the objection that becomes apparent when we adopt a votercentered perspective.Since it promotes conventional representative systems whose central institution is elections, the objection entails that the voters' lack of accountability would only be fatal when they participate in one kind of mass votes-popular votes, in which ballots display policy options-but not in the other-elections, in which ballots display candidate or party options-even though voters do the same thing in both cases-cast a ballot.Proponents of the accountability objection justify this distinction by suggesting that voters only need to be accountable when they cast their ballot in popular votes because referendum and initiative processes would transform them into 'legislators'.A more formal reconstruction of the accountability objection thus looks as follows: My argument highlights two weaknesses of this objection in turn.The first one is that the scope of the objection is broader than anticipated by its proponents: the accountability objection cannot be consistently used to dismiss popular vote processes without being also fatal to elections (Sect.'The Scope of the Objection: Voters in Popular Votes and Voters in Elections').Disentangling two different meanings attributed to the term 'legislator' when it comes to qualifying ordinary voters as such by proponents of the objection, I demonstrate that voters 'legislate' either 4 Some versions of the accountability objection are softer, which contend that the lack of accountability of voters in popular votes might not be a reason not to implement them, but that their especially 'high cost to retrospective accountability' ought to be acknowledged (Warren 2014, p. 46).While this article focuses on the stronger version of the objection, I take the arguments presented here to question these softer versions as well. 5In order to stay focused on this accountability objection, I will also not question the premise that elected actors can be held accountable in appropriate ways (but see, e.g., Warren (2014, p. 45) for nuances of this view).
in both kinds of mass votes or in neither (and so the distinction between P2 and P3 collapses).The second weakness is that the objection wrongly characterizes voters as the kind of 'legislators' who should be accountable (actors to whom P1 applies) (Sect.'Rescuing Mass Votes From the Accountability Objection').I argue that this is because, even if voters are formally empowered to participate in lawmaking, they do not act on behalf of others in ways that justify demands for accountability.This does not exclude that voters have responsibilities when it comes to casting a ballot.By way of closing the article, I thus introduce some reasons why, compared to conventional representative systems, democratic systems including well-designed popular vote processes could offer more supportive conditions for voters to meet such responsibilities.
These considerations contribute to reinforcing the case for democratic systems that include popular vote processes by addressing one objection to them that has received little attention.More broadly, the article contributes to clarifying the institutional role and power of ordinary voters in all kinds of mass votes in democratic systems.Doing so also highlights a useful conceptual distinction between two meanings of 'legislator' (as representative and as lawmaker) often conflated in democratic theory, and questions problematic assumptions about citizens having unmediated decision-making power in 'direct democracy'.These clarifications may invite reconsiderations of other objections to popular vote processes and help in developing new insights to determine what, if anything, should be expected from citizens in democratic decision-making processes.

The Scope of the Objection: Voters in Popular Votes and Voters in Elections
Voters' lack of accountability is considered an important objection to referendum and initiative processes because, unlike in elections (P2), voters in popular votes would be 'legislators' (P3) (of the kind who need to be accountable (P1)).The distinction between P2 and P3 is essential to consistently argue that the accountability objection applies to democratic systems that include popular vote processes, but not to conventional representative systems.It reflects on a long tradition of conceiving of all kinds of binding referendum and initiative processes as processes fundamentally distinct from electoral processes and that empower voters to 'govern' or to 'directly legislate' (see e.g., Magleby 1984;McKay 2018).But in what sense, exactly, are voters 'legislating' when they cast their ballot in popular votes-and not when they vote in elections?
In this section, I disentangle and assess two conceptions of voters in popular votes as 'legislators' proposed in the literature: as lawmakers and as representatives.6I suspend judgment concerning whether voters should be conceived as lawmakers or as representatives, and whether these 'legislators' should be held accountable, until the next section.Rather, my aim here is to demonstrate that neither conception of voters as 'legislator' provides a sound basis for preserving a strong dichotomy between voting in popular votes and in elections.This highlights the first weakness of the accountability objection to popular vote processes: its scope is wider than anticipated by its proponents.It cannot be an objection only to popular vote processes; it must be an objection to all mass vote processes.

Voters in Popular Votes as Lawmakers
A very widespread way of understanding voters in popular votes as 'legislators' emphasizes the lawmaking competences formally granted to them.Voters in popular votes would be 'legislators' because they have the 'ability to make coercive law that binds others' (Serota andLeib 2013, p. 1605; see also Hamilton 1997, p. 7). 7 This special ability would derive from two key differences between voting in popular votes and voting in elections, which I discuss in turn in this sub-section: only voters in popular votes would make decisions that are not mediated by the filter of representative institutions; and only voters in popular votes would make decisions on policy issues.

Filter-Less Lawmakers
A common defense of conventional representative systems insists that elections ensure that voters' impact on policies is filtered.Voters' possibly self-interested, uninformed, or prejudiced electoral decisions may impact which candidates are elected; but only those elected would then be able to make laws-and they would do so while being accountable to the public, which would incentivize them to 'take into account many views including the larger public good' (Hamilton 1997, p. 7).In contrast, popular votes would display what Serota and Leib call a 'lack of filtering', whereby voters could cast their ballots directly on policies and, thus, make themselves the law (Serota and Leib 2013, p. 1611, also p. 1604; see also Hamilton 1997, p. 13).
The key problem with using the absence of filters to distinguish voters in popular votes as 'legislators' from voters in elections is that in political practice, the ballots of voters in popular votes are very much subject to the filter of electoral representative democracy.First, casting a ballot in a popular vote involves voters choosing between pre-selected options-generally binary, yes/no options to select either the status quo or a proposition for change-which are greatly shaped by elected actors.This is especially the case in referendum votes.These include mandatory referendums (which elected actors need to organize on certain pieces of legislation they adopted in order for them to be approved); top-down referendums (which elected actors can organize on a proposition of their choosing); and bottom-up referendums (which nonelected actors can demand on laws adopted by elected actors).For all these processes, voters can choose between the status quo and policy propositions that were previously endorsed and adopted by elected actors-usually, following lengthy processes of deliberation and decision-making in electoral spheres, which must meet the relevant legal, constitutional, and procedural requirements.
Electoral filters on ballot options are less strong in popular votes on initiatives, where the propositions opposed to the status quo on voters' ballots originate from nonelected groups.But electoral filters are not absent either.One thing to clarify is that despite the widespread notion that initiatives processes empower ordinary voters to make their own propositions, the institutional role of initiative 'sponsors' or 'committees' is distinct from that of voters.Only specific groups-often composed, at least in part, of elected officials or party members (see, e.g., Gerber et al. 2001)who succeed in collecting a sufficient number of signatures from their fellow citizens within a certain period of time in support of their propositions formally receive rights to determine whether to maintain their proposition and place it in on the ballot, as well as to campaign for it in the media or in official information documents (see el-Wakil 2020).In contrast, ordinary voters' only prerogative in initiative votes is to select among pre-defined options-which were filtered by these groups and by the signature collection process.8Furthermore, beyond these nonelectoral filters, electoral ones are also at play.In some cases, as for instance in Switzerland, elected members of parliament must validate initiatives after the collection of signatures, and they can propose a counterproposal to an initiative.This counterproposal is then either also put to the vote, as a third alternative, or adopted in case initiative sponsors decide to withdraw their proposition (Lacey 2014;McKay 2018).
A second important filter of electoral democracy comes after the popular vote, for both referendums and initiatives.When propositions for change are adopted by a majority of voters, elected actors generally have (and use) the competence to create the specific regulations that will implement this decision and to devise how, and when, it will be put in effect (see, e.g., Gerber et al. 2001).Moreover, if voters in popular votes may 'make binding decisions [elected actors] disapprove and dislike' (Auer 2016, p. 402), the same is true of voters in elections, whose decisions can even eject these representatives from office.
The notion that voters in popular votes have an unfiltered impact on policy decisions compared to voters in elections thus mischaracterizes the practice of popular vote processes.It also reinforces a misleading interpretation of referendums and initiatives as processes that grant unmediated lawmaking power to 'the people', thus diverting us from acknowledging the major role played by elected actors in formally determining popular votes' timing, options, and effects-an issue to which I return below (Sect.'Voters as Co-Lawmakers').For now, what I have proposed is that we cannot distinguish voters in popular votes from voters in elections because the former would make laws without the filter of conventional representative institutions.

Lawmakers on Policy Issues
We could still understand voters in popular votes as 'legislators' in the sense that they have the competence to participate in making coercive decisions on policies, whereas voters in elections only participate in making decisions on people (Serota andLeib 2013, p. 1598).Yet I argue that this criterion also relies on too strong a dichotomy between voting on people in elections and on issues in popular votes to justify the distinction between P2 and P3.
On the one hand, and even if they are formally about policies, popular votes as they occur in contemporary democratic systems are often also de facto votes on people.This is most obviously the case when elected actors tie their mandate to popular vote's outcomes, as in the 2016 constitutional reform in Italy where then Prime Minister Renzi announced that he would resign if the reform was rejected (and did so after the popular vote).9Even when they do not drive elected actors out of office, popular votes have been acknowledged as a way for voters to hold elected actors accountable while in office (in addition to elections) (Setälä 2006;Lacey 2017;Cheneval and el-Wakil 2018).In referendums, voters can impose clear consequences onto elected actors by accepting or rejecting the policy propositions they developed and adopted.In both referendums and initiatives, voters' ballots can also affect the level of support of specific parties or elected officials.Voters can decide to provide or withdraw support from a party or an elected official's popular vote campaign by following or rejecting their voting recommendation; and it has long been established that, in the Swiss and US context, a significant percentage of voters rely on partisan cues to make their voting decision, thus voting not only on the issue at stake but also in support or against a party (see, e.g., Lupia and Matsusaka 2004;Colombo and Kriesi 2017).Popular vote decisions can thus not only influence which policy propositions are adopted, but also which people gain or lose support.
On the other hand, voters in elections may not only select representatives, but also policy propositions.Democratic theorists have long questioned the democratic character of the elitist account of elections, according to which voters' only role is to select representatives based on who they are and let them govern until the next election (Schumpeter 2008, pp. 269-273).Contemporary representation theories insist that some level of responsiveness to citizens' policy preferences is one of the core democratic features of representative governments (see, e.g., Pitkin 1967;Urbinati 2006;Lafont 2020): 'when we vote [in elections], we do two things at once: We contribute to forming a government or opposition, and we seek representation of our positions and preferences' on issues (Urbinati and Warren 2008, pp. 398, 402). 10uerrero similarly recognizes that elections can provide elected actors with specific policy mandates (Guerrero 2010; see also Goodin and Saward 2008).In this perspective, elections are considered to allow voters: to choose the persons who will represent them, but also to give these persons mandates to advance the policies and ideas they propose or represent; to influence MPs on their policy decisions between elections, by anticipation of the next election; and to decide whether to re-elect them at this next election based on their retrospective appraisal of the performance of their representatives, also in terms of their capacity to deal with specific issues (Goodin and Saward 2008;Urbinati and Warren 2008;Mansbridge 2009).
By casting their ballot, voters in both elections and popular votes can (and should) thus participate in making binding decisions that influence both people and issues.While the specific impact of these collective decisions on society as a whole in terms of effecting change varies from case to case, any mass vote can be consequential in terms of approving specific laws or policies, of granting or removing support to the representatives who designed and campaigned for them, of selecting who will run the country for the next three to seven years, or of giving these actors policy mandates.And just as '"the people" who votes a policy into law has no accountability for the constraints on executing the policy or for its consequences' (Warren 2014, p. 46), 'the people' who elect representatives cannot be held accountable in the sense of being subject to potential consequences for its decision.

Voters in Popular Votes as Representatives
A second, less common way to understand voters in popular votes as 'legislators' should be mentioned, which is proposed by Serota and Leib: unlike voters in elections, voters in popular votes would be representatives (of a nonelected kind (see, e.g., Montanaro 2012)).In their view, 'citizens voting in direct democracy are representative legislators', whereby 'voters in direct democracy are in a relationship of political representation with the citizenry' (Serota and Leib 2013, pp. 1596, 1597).More specifically, they write (Serota andLeib 2013, p. 1605): That the citizen who casts her ballot in a direct democracy election represents others is easy to perceive under a range of circumstances: win or lose, the voter is pulling a lever for the too-young, too-infirm, too-lazy, and too-felonious.
Whether a meaningful relationship of representation is at stake here, and whether it justifies demands of accountability, will be discussed in Sect.'Rescuing Mass Votes From the Accountability Objection'.For now, I only wish to emphasize that this criterion cannot justify a distinction between voters in popular votes (P3) and voters in elections (P2).Voters in elections also 'pull a lever' for those who do not cast a ballot by deciding which candidates access elected offices; and electoral votes generally involve the same exclusions from universal suffrage as popular votes, so that voters in elections would also be representing those without a right to vote.Therefore, a conception of voters in popular votes as 'legislators' in the sense that they would be representatives fails to justify the distinction between 'legislating' voters in popular votes and 'non-legislating' voters in elections.
Thus, the accountability objection's central distinction between P2 and P3 cannot be upheld: none of the reasons offered for considering voters in popular votes, but not voters in elections, as 'legislators' appears to withstand scrutiny.As a result, it is inconsistent to reject only popular vote processes while preserving elections.Voters' lack of accountability is either fatal to both kinds of mass voting processes, or to neither.

Rescuing Mass Votes From the Accountability Objection
Which is it?As the revised formulations of the objection below suggest, this depends on whether or not ordinary voters are 'legislators' who need to be held accountable (version A) or not (version B).In this section, I begin by presenting the justification for P1 in democratic theory.This serves to clarify who are the 'legislators' who should be held accountable, in the sense of having to face consequences imposed by others for their actions and decisions, in democratic systems.I then revisit the two conceptions of voters as 'legislators' introduced in the previous section-as lawmakers and as representativesin this light to examine whether they justify considering voters' lack of accountability as a reason to dismiss mass votes.This discussion offers additional support to the view that voters do not need to be accountable in democratic systems (see Ferejohn 2008;Lever 2015;Lafont 2020) and highlights the second weakness of the accountability objection by demonstrating that voters have lawmaking powers but that their powers are not exercised for others in relevant ways.

Why Should 'Legislators' be Accountable?
Proponents of the accountability objection take it as granted that 'legislators' should be accountable.But why would certain actors have to face possible consequences for their actions or decisions?A general answer to this question found in democratic theory is that accountability is required when certain actors are entrusted with political powers to act on behalf of others.As Warren (2014, p. 39) writes, it is 'the condition of vulnerability to power [that] creates the need for democratic accountability '. 11 He adds that in a democracy the most basic problem to which accountability is the solution is that even though government is 'owned' by the people, the powers of collective decision and action are held by elites, to which the people are vulnerable.(Warren 2014, p. 41) In this view, accountability serves to democratically legitimize unequal provisions of formal political power by serving two purposes.
On the one hand, accountability provides a way to control powerful actors to those who are impacted by these uses of power. 12The ones vulnerable to power, on whose behalf actors entrusted with special powers act and whose use of power coerces them, should have the possibility "to" demand justifications and impose sanctions onto those in power in ways that can limit or interrupt possible abuses of power (Urbinati and Warren 2008, p. 402;Warren 2014, p. 40).
On the other hand, accountability sets incentives for powerful actors to behave well.This is generally understood to mean that they should represent, in a normative sense, those impacted by their use of power (see, e.g., Pitkin 1967;Manin et al. 1999;Urbinati 2006).Here, the possibility to face consequences imposed by those vulnerable to power is thought to increase the chances that actors in power act as good representatives, namely foster the interests of those on whose behalf they act and stay responsive to their demands (see, e.g., Manin et al. 1999;Mansbridge 2003;Warren 2014, p. 39).The concern that voters as 'legislators' will not behave well is, as briefly mentioned in the introduction, a central worry of proponents of the accountability objection.For instance, Hamilton argues that without accountability, the incentive for deliberation experienced by representatives disappears [for voters].The absence of accountability feeds ignorance and self-deception, which in turn reduces the likelihood that any individual voter will be competent to vote on the issue [and will] take into account many views including the larger public good.(Hamilton 1997, pp. 12, 7).
Similarly, Landa and Pevnick insist that 'by linking continued office-holding to public approval, representative systems generate incentives for officeholders [who desire to retain offices] to exercise power with due consideration for the likely effect of policies on the welfare of ordinary citizens' (Landa and Pevnick 2020, pp. 4-5).Such incentives 'will lead representatives to exercise political power more responsibly than would citizens in a direct democracy' (Landa and Pevnick 2020, p. 4).As for Serota and Leib, they propose, at the end of their article, that the secret ballot should be abandoned, so as to set the precondition for voters to potentially have to face consequences for their votes-which would in turn increase the chances that voters use their power to 'do the right thing' (Serota andLeib 2013, p. 1615).
But do voters really belong to the kind of actors who should be accountable, in democratic systems?The justification of accountability summarized here specifies the relevant meaning of 'legislators' in the first premise of the accountability objection (P1): democratic systems should guarantee that the actors entrusted with political powers to act on behalf of others can be held accountable.Are ordinary voters granted political powers to act on behalf of others?I consider this question in what follows by critically revisiting the two conceptions of voters as 'legislators' introduced in Sect.'The Scope of the Objection: Voters in Popular Votes and Voters in Elections'-as actors entrusted with lawmaking powers and as representatives acting on behalf of others.

Voters as Co-Lawmakers
Are voters formally entrusted with lawmaking powers, as suggested by proponents of the accountability objection?I demonstrated above that they do not have powers to make laws without filters.I also explained that voters' lawmaking competences are very limited-especially compared to the competences of elected actors.To put it somewhat schematically, elected actors are formally entrusted with the full range of competences associated with lawmaking (McKay 2018, p. 332)-they can set the agenda, draft and develop laws and policies, deliberate, find compromises, and make decisions on these laws and policies and on their implementation for the entire political community and for the entire duration of their term.If top-down referendums are available, they can call for a popular vote on a question of their choice; when elections can be called by elected actors, they can select the date of the vote.And, beyond these formal competences, they also generally play a large role in framing mass votes' campaigns due to their privileged access to the media.In contrast, voting rights provide citizens with the competence to cast a ballot in punctual mass decision-making moments to select among pre-defined options.Acknowledging this point is essential to undercut populist interpretations of the results of both popular votes and elections as 'decisions of the people', reflections of the 'popular will', or instances of 'the people governing' for which representatives should not be held accountable (as argued, e.g., by Setälä 2006;Weale 2018).Elected actors' profound impact on the results of mass votes should be acknowledged.
Yet doing so does not entail that voters have no political power.A key characteristic of voting in democratic systems is that it should be consequential, as punctual moments of decision-making open to the entire electorate, with universal suffrage serving as a fundamental way in which empowered inclusion is realized (Warren 2017).Many have nonetheless argued that voters have no causal power over the outcome of the vote because it is extremely unlikely that voters cast the decisive ballot, the one ballot that makes the outcome be a yes or a no, party A or party B (see, e.g., Beckman 2017).But this take displays a reductive understanding of what matters, in the result of a vote, in terms of impacting the entire political system over some time.Beyond which decision is made, how this decision is made matters, too: voting measures 'the strength of a decision' and displays how large the opposition is (Warren 2017, p. 49).For Guerrero, how electoral decisions are made affects the normative mandates of elected actors by recording their degree of support (Guerrero 2010, pp. 274-277).And it has also been shown that, for instance in US states, 'larger margins of victory were associated with fewer attempts at legislative alteration [in the implementation of an initiative], while narrow margins of victory generated more frequent, and more extensive, attempts at alteration' (Ferraiolo 2023, p. 26).More generally, both the level of turnout and the relative sizes of majorities or pluralities affect who gets into office, whether parties will be able to decide on their own or whether they will need to compromise or make alliances with other ones (and which ones), whether a proposal rejected in one vote will be modified and put to a popular vote again or whether it will be abandoned (at least for a while).
If we accept that formal lawmaking powers come in degree, and to reflect the fact that voters are not the only institutional actors making laws, I thus propose to extend Lafont's (2020, p. 202) characterization of voters' role in popular votes as that of 'co-legislators' to voters in general-in the sense that they are 'co-lawmakers'.Voters' formal competences are very limited compared to other co-lawmakers, such as elected authorities.But these powers still enable them to participate in making decisions that impact the entire community-both in terms of which policies will be implemented and of who will gain mandates to rule.

Voters Do Not Represent Others
Recognizing voters' co-lawmaking powers is, however, not sufficient to establish that voters should be held accountable.For this claim to be sustained, what needs to be demonstrated is that voters exercise this power on behalf of others, who are vulnerable to their use of power.I thus return to the second conception of voters as 'legislators' discussed in Sect.'The Scope of the Objection: Voters in Popular Votes and Voters in Elections', according to which voters in popular votes 'are structurally in a relationship of political representation with [their] fellow citizens' (Serota andLeib 2013, p. 1604) and act on behalf of those who do not vote-either because they do not turn out or because they do not have the right to vote.I explained above that this is the case for voters in elections as well.But are voters structurally exercising co-lawmaking power as representatives?Voters of course lack common defining characteristics of representatives, such as being selected or authorized by others (Mansbridge 2009), making claims to represent others (Saward 2010), or being acknowledged as standing in for others to do certain things (Rehfeld 2017).What Serota and Leib seem to have in mind, instead, is that voters represent others precisely in the functional sense of acting on their behalf.In support for their claim, they (Serota andLeib 2013, p. 1604) refer to the works of Stephan (2004) and Warren (2008), whose contributions emphasize the role that ordinary citizens can play as representatives.
For Warren, it is randomly selected participants in citizens' assemblies and other kinds of mini-publics who are 'citizen representatives' (Warren 2008, p. 57) in the sense that they are entrusted with powers that others do not have (and should thus be constrained by (nonelectoral) accountability mechanisms).Indeed, access to citizens' assemblies and the competences granted to them is restricted to those who are selected: 'because most citizens do not participate through these institutions […] we should be conceiving of them [Citizens' Assemblies] as representative bodies' (Warren 2008, pp. 56-57).However, the extent to which mass votes also place ordinary voters in representative roles is doubtful.Access to voting is not restricted in the same way as access to mini-publics: whereas, in assemblies, the selected few act on behalf of all who are not included, a core characteristic of mass voting processes is precisely that all of the electorate's members can use their ballot and vote for themselves.Stephan (2004, p. 119) proposes to understand citizens as representatives when they voluntarily become 'active in the public policymaking and implementation' and create organizations.They should in this view again be understood as representatives because they act for others in the sense that they 'do something we cannot do ourselves' because they have some special (formal or informal) competences or resources (Stephan 2004, pp. 120, 124).Thus, again, this characterization of citizens as representatives acting on behalf of others does not apply to participating in mass votes: voters all have a ballot, and can decide to cast it or not.
Therefore, the claim that voters exercise co-lawmaking powers on behalf of their fellow citizens who do not turn out must be rejected.Even when those who turn out collectively make decisions that will also impact those who do not turn out, the fact that their co-lawmaking powers are equally accessible to those who do not turn out entails that they are not acting on behalf of non-voters. 13As a result, voters need not be accountable to non-voters.
Are voters exercising their co-lawmaking power on behalf of those who are not enfranchised but who are impacted by these uses of power?In this case, voters' right to participate in making decisions is a power to which those who do not have a right to vote do not have access.Yet requiring that voters be accountable to the non-enfranchised would appear as a weak treatment for the symptoms of a bigger problem that affects the legitimacy of democratic systems more broadly, namely that of undemocratic exclusion.Even if voters were accountable to non-enfranchised people, the latter would still be vulnerable to the much more extensive powers of elected actors and lack resources to hold them accountable.Expanding empowered inclusion is likely to be the only remedy to this problem.
The conclusion of this section is thus that voters as co-lawmakers do not belong to the category of 'legislators' who should be held accountable.Both elections and popular votes are rescued from the accountability objection, as summarized in the reformulated objection below.

The Potential of Democratic Systems with Popular Vote Processes
That ordinary voters in democratic systems do not need to be accountable, in the sense of having to face consequences imposed by others for their voting decisions, does not entail that they have no responsibilities or face demands for other forms of accountability related to their co-lawmaking powers.For instance, some have argued that the deliberative ideal of democracy requires voters to be accountable in the sense that they should base their voting decisions on 'reasons that all citizens can reasonably be expected to endorse' (Engelen and Nys 2013, p. 495).Others have proposed that voters should gain information to hold elected actors accountable and anticipate the consequences of voting decisions (Warren 2014, pp. 45-46).It is not the purpose of this article to determine what the responsibilities of voters are (or whether they have any).But, to not leave concerns about uninformed and prejudiced voters unaddressed-which, as mentioned above, are a key concern of proponents of the accountability objection-and by way of bringing this article to a close, I wish to briefly highlight three reasons why, against the accountability objection to popular vote processes, systems with well-designed popular vote processes could be expected to offer voters more supportive conditions to meet such demands than conventional representative systems.First, democratic systems with popular votes can enhance voters' provision of relevant information, especially in the mid and long term, with more frequent mass campaigns that discuss more diverse issues.In both electoral and popular vote campaigns, elected and nonelected actors have an interest in engaging in mass debates and in reaching out to the public with arguments and voting recommendations to convince majorities of citizens of their favored candidates or policies (Lacey 2017).Including popular votes in addition to elections can also serve to focus the discussion on specific issues, thus enabling voters to learn about elected actors' ongoing decisions (with referendums) and positions on specific propositions (with both referendums and initiatives).Popular vote campaigns also have the advantage of structuring the public discussion between 'yes' and 'no' campaigners, without necessarily reducing the variety of arguments proposed in favor of and against the contested law or policy (see Cheneval and el-Wakil 2018, p. 298).As such, democratic systems with popular vote processes (and well-designed campaign regulations (Suiter and Reidy 2020;Renwick et al. 2020)) can provide citizens with information about candidates, about representatives' favored policies and agenda-setting intentions, and about policies and the arguments surrounding them.
A second way in which democratic systems with popular votes could be expected to increase voters' information is that regular and diverse votes can emphasize the ways in which mass vote decisions can impact voters' daily lives (see Cheneval 2015, p. 167), and thus motivate them to get informed.Popular votes' outcomes are not always immediately tangible, as for instance with popular votes on tax reforms.But, especially in systems with proportional elections where the effects of voting decisions might be hard to experience in the short term, having also votes on the retirement age, an additional week of vacation per year, or the legalization of samesex marriage does demonstrate that voting matters.Not all voters have to bear the same consequences, in both elections and popular votes.Yet adding popular vote processes to democratic systems increases the chances that most voters will be impacted by mass votes at some point-especially in societies structured by crosscutting cleavages.While this would need to be assessed empirically, systems that include popular vote processes could help voters to see the point of getting informed before heading to the voting booth.This is, of course, not to say that systems with popular vote processes are a panacea when it comes to enabling voters to cast informed ballots.The unequal circumstances of politics remain, in which various actors can try and take advantage of the opportunity structure, possibly at the expense of providing accurate or complete information, and other actors might not succeed in gaining visibility.The extent and quality of information also depends on the quality of the media system, the engagement of nonelected actors, and, as mentioned above, the specific regulations of electoral and popular vote campaigns.Efforts to rebalance discursive inequalities and undermine the success of misinformation strategies are required, which might include inviting mini-publics to examine the claims proposed in mass campaigns and write accessible summaries of relevant arguments for fellow voters (Citizens' Initiative Review (Gastil and Knobloch 2020), which could be adapted to electoral campaigns, too).But we should also acknowledge that compared to democratic systems that only empower voters to cast their ballots in elections, democratic systems with popular vote processes set a basic institutional opportunity structure that can provide voters more readily with relevant informational resources and with additional incentives to get informed.
Still, this optimistic outlook might not convince the skeptics of popular vote processes or of voters' competences.For them, a final point to emphasize is that systems that include two kinds of popular vote processes, mandatory and bottom-up referendums, offer institutional resources to limit the possible damage caused by 'irresponsible' voters.On the one hand, these kinds of referendums only empower voters to vote on whether to introduce law propositions that have been considered acceptable by elected actors.On the other hand, systems with mandatory and bottom-up referendums limit the consequentiality of electoral votes: in such systems, elections only provide elected actors with a conditional authorization to make all of the decisions on the issues for which they are competent unless a popular vote is required or demanded.Thus, whereas conventional representative systems offer no consequential ways for voters to correct their electoral decisions in light of new information before the next election, systems with these kinds of referendums therefore provide opportunities to challenge and possibly cancel certain decisions of elected actors during their time in office.

Conclusion
The accountability objection to popular vote processes contends that binding referendum and initiative processes should not be introduced in democratic systems because they empower voters to 'legislate', yet offer no way to hold them accountable, in the sense that no consequences can be imposed onto voters by others.Therefore, democracy would be better realized through conventional representative systems.In this article, I argued that this objection has two major weaknesses.First, it is inconsistent: the objection cannot be simultaneously used to reject popular vote processes and to endorse elections as a central process of conventional representative systems.This is because we lack sound justifications for characterizing voters in popular votes, and not voters in elections, as 'legislators'.This conceptual clarification may have implications for other objections to popular vote processes, such as the contention that voters are not sufficiently competent to cast ballots on policy issues.Second, the objection mischaracterizes voters as the kind of 'legislators' who ought to be accountable.I advanced that, even if we acknowledge the co-legislative role played by voters in lawmaking processes, their powers are not the kind of unequal, special powers that require accountability.I closed the discussion by suggesting that including well-designed popular vote processes in democratic systems could well set an opportunity structure that supports voters in gaining the information which they need to play their institutional role better than in conventional representative systems-or, at least, enable them to correct for previous voting decisions in light of new information.While this remains to be further assessed empirically, the answer to the accountability objection proposed in this article can hopefully serve to reconsider the democratic potential of the decision-making methods put in place in political systems that include referendums and initiatives.