Pitfalls on the Road to Frey’s Democracy of the Future Comment on ‘‘Proposals for a Democracy of the Future’’ by

In this article, we discuss some of the recent proposals developed by Bruno Frey for making future societies more democratic and prosperous. While sharing Frey’s conviction that new ways of involving citizens in political decision-making in representative democracies are needed, we engage in debate on his proposals on voting rights, referenda and randomized decision-making. We are critical of the random selection of ofﬁce-holders proposed by Frey, suggesting it may lead to alienation by citizens from the political process and would not really address the perceptional gap between elites and the broader public. Mindful of the pitfalls inherent in the increasing use of referenda, we suggest pre-referendum deliberation rather than the post-referendum conciliation proposed by Frey. Fur-thermore, we believe that it is almost impossible to propose good solutions without engaging the different arguments and empirical ﬁndings exploring the substantive as well as the procedural causes of democratic discontent.


Introduction
In his article 'Proposals for a democracy of the future' Bruno Frey (2017) develops a number of ideas and specific proposals for making future societies more democratic and prosperous. He argues that direct citizen participation makes democracies stronger and citizens more satisfied and outlines how participation can be improved via nine different routes. His proposals include, among others, introducing weighted voting rights based on length of residence, holding referenda that include post referenda conciliation procedures, and using random procedures in elections for public bodies.
We share Frey's conviction that there is a need for new ways of involving citizens in political decision-making in representative democracies. Many of his proposals break new ground, while developing themes which have been prominent in his work. Next to their original contribution, these proposals represent an opening for further debate. We would like to enter this debate and discuss some of the potential pitfalls and limitations of the proposals. We also suggest some alternative ways to advance citizen participation.
We agree with Frey's observation that recently we have witnessed unprecedented attacks against existing democracy. We would go even further by claiming that we have entered a period of discontent with democracy in its current form which affects even the most stable and developed democracies in the world. While for Frey the exploration of the possibilities of direct participation is a long-lasting research interest, current developments in the world make his contribution particularly interesting and topical. The question is, however, whether the remedies that are offered are the right medicine for the current malaise of democratic discontent. We start with some general observations before discussing several of Frey's proposals in more detail.
A first general observation we would like to make is that a contribution based on public choice and limited to institutional adjustments may leave some of the most important causes of current citizen discontent, based on actual outputs of democratic polities, unrecognized and unacknowledged. It suffices here to mention debates about rising inequality and its effects, as exemplified in the much-discussed work of Piketty (2014) or the increasingly rich and nuanced debates about the effects of globalization in creating new cleavages between cosmopolitan and nation bound citizens (Kriesi et al. 2006(Kriesi et al. , 2008. Naturally, debates about these causes are far from conclusive. Nevertheless, we believe that a discussion of possible changes in the organization of decision making and citizen inputs in democracies should take into account the substantive causes of the 'democratic discontent' or the discontent with democracy. Identifying the factors that have made developed democracies highly vulnerable to discontent at this point in history remains a difficult task and a long process of scientific discovery. Debates about causes can be strongly politicized and influenced by the values and preferences of observers. Frey avoids this trap by not taking much of a stand on the causes of the current crisis of democracy in substantive terms, but this leads to his core assumption that citizen preferences are procedural rather than substantive. Yet the assumption that public discontent is rooted in a lack of voice and influence over decision making rather than in the substantive outputs of decision making in currently existing democracies needs to be examined more carefully and supported by more data than he currently presents. We question whether the possibility of (more) voting can make people more satisfied: it is possible that the increased satisfaction of migrants when they become Swiss citizens, which Frey cites as an example of the increased participation effects on happiness, could have other causes, for example the improved access to public goods, services and to job opportunities which citizens have in comparison to non-citizens.
A second observation is that public choice analyses sometimes tend to build on existing governance structures that are treated as 'given' in the analysis. One should not forget that the very institutions that make collective decision-making possible are based on political choices as well, which makes these institutions, to some extent, endogenous to the analysis. Here we encounter the well-known problem of institutions as 'relative' constants, extensively discussed by Riker (1980). Introducing political costs to decrease the likelihood of changing institutions as compared to ordinary political decisions creates stability. Another way of strengthening this relative stability is by focusing on the underlying values in society that form the foundation of democratic systems. Commonly shared understandings about the desirability of democracy as a political system, legitimacy of its institutions and decision-making procedures, form a stable foundation for collective decision-making. This, however, requires a continuous confirmation of societal consensus and the shared core values in daily politics. It is an open question whether some of the suggested solutions will indeed, as intended, strengthen democracy and improve citizen participation, or could, potentially, contribute to a gradual erosion of some of these values.
Following our general observations, we now discuss some of Frey's specific proposals on voting rights, referenda and randomized decision-making.

Voting Rights
The rising number of studies showing the emergence of a new cleavage between mobile citizens and those whose rights and livelihood are protected by national borders also discuss the transformation of political landscapes in Western democracies linked to this development (Kriesi et al. 2006(Kriesi et al. , 2008. Frey's proposal for having a voting share proportional to time spent in a jurisdiction-although perhaps difficult to administer-allows for some redressing of the imbalance between mobile and less mobile citizens in terms of the rights and public goods they enjoy. This proposal may prove tremendously controversial for cosmopolitan citizens of Europe or the world, but might, if implemented, address the problem of voters influencing polities and policies away from the place where they actually live and pay taxes. In our view, Frey takes a 'material' instead of 'formal' perspective on citizenship and public participation by making voting rights dependent on the time spent in residence. We agree that in our increasingly more mobile societies in which individuals from other countries live, work and benefit from local and national public goods and services, a formal perspective on voting rights does not suffice. It deconstructs the logical triangle between enjoying public goods, the obligation to pay for these through taxation and the possibility (e.g. right) to decide on different quantities, if one wishes. The latter, however, is, at least in the European Union, only made possible in local and European elections. For national elections 'formal' citizenship is still key. So, it is no surprise, that from an economic perspective and possibly also from a normative one, this arrangement is no longer optimal. By giving all users of public services a say in its production through elections, they all may be better off. This would indeed require a change towards 'material' citizenship as Frey proposes.
The same logic can be applied to the position of another group of residents, who may have too generous rights. These are the holders of double citizenship: people who, for whatever reason, have double passports. Mostly these are citizens who have lived for a long period in one country and obtained citizenship, but are also still citizens of another country based on national law. To put it differently, these residents have 'formal' and 'material' citizenship in their country of residence and still 'formal' citizenship in another country. These voters may have little connection with the political developments and discussions in the country where they do not live but can vote. It does not seem reasonable for them to be still in a position to determine public policy in their original country of citizenship. The diaspora may even be asked to vote on the shape of the polity, as in the case of constitutional changes proposed in the Turkish constitutional referendum of April 2017. If a citizen is not likely to experience the effects of decision making in a political system or does not pay (full) taxes, why should this person be in a position to determine policies or institutions?
One pragmatic way of introducing weighted or material voting rights to European citizens would be to connect such proposals to the EU's Residence Directive (2004/ 38) in which residency rights of European citizens in other EU countries are arranged. Basically, after 5 years of living in another EU country, EU nationals acquire the right of permanent residence. This right should also involve voting rights, in the country of residence-at the expense of the voting rights in the country of origin. Based on the already existing legal framework, having only voting rights as a permanent resident in another EU country might be feasible. A serious drawback of such a change, however, would be the complexity involved in securing adequate representation based on residence: tracking how much time a citizen spends in a specific jurisdiction. The arrangements for non-EU citizens would be even more complicated, as they would require bilateral arrangements between the member states of the EU and other countries that also may have an interest in moving to 'material' forms of citizenship. Finally, tracking how much time citizens spend in one jurisdiction as opposed to another may in itself have adverse effects on personal freedom and thereby on citizen satisfaction with democracy.

Referenda
Referenda have long been seen as an obvious tool to get citizens involved in policymaking, especially in representative democracies. In his work, Frey has often recommended referenda as a way to involve citizens in key decisions of their polity, not only to articulate voter preferences, but also ''… to break up the politicians' coalition against voters'' (Frey 1992, p. 218; see also Frey and Bohnet 1993, p. 72). Interestingly, since then, many countries have held ad hoc consultative referenda on important political decisions or even introduced legislation allowing citizens to initiate referenda. For example, in 2003, eight candidate countries from Central and Eastern Europe organized referenda asking citizens to approve their accession to the EU, which resulted in overwhelming support to the 'yes' camp. 1 There are also less glorious examples, among them the recent Dutch referendum on the European Association Agreement with the Ukraine which we have closely observed. Held in April 2016, this referendum had a turnout of 32% of the voters, 61% of which voted 'against' the ratification of the agreement. 2 Although the referendum was consultative, after the vote a majority of members of parliament indicated that they could not ignore the result, while initially they had voted to ratify the agreement. Eventually, under high international pressure, Dutch Prime Minister Rutte negotiated a declaration added to the Agreement in which some of the concerns raised during the referendum campaign were addressed (e.g., the so-called membership perspective for Ukraine). Having adopted the declaration at an EU summit, it was possible to find a new majority in the second chamber in favor of ratifying the agreement. 3 Following these experiences with the first application of the Dutch law on referenda, the State Council, an important advisory body for the Dutch government, warned about the polarizing effect of referenda. The advice highlighted the tension between this instrument and the rule of law (Raad van State 2017, p. 17).
In our view referenda are, in general, a rather crude way of identifying what voters like. As explained elsewhere (Steunenberg 1992), 'yes-no' type of voting prevents voters from expressing any view more nuanced than the answer to the question on the ballot. Furthermore, the possibility that others may like a 'third' option, not included in the take-it-or-leave-it vote, may lead to the alienation of that group of voters. Some voters, simply, may not want to vote since they do not like either option. That was arguably the case with some of the Dutch voters who 1 Ranging between 66.8% approval in Estonia to 93.7 in Slovakia and 91.9% in Lithuania. Malta also held a referendum in 2003 in which 53.6% of citizens voted for accession. Later, in the second wave of the Eastern enlargement, Bulgaria and Romania refrained from putting the decision to accede to the EU to the popular vote. 2 Survey research after the referendum (Jacobs et al. 2016) identified the main concerns among the opponents to be the (perceived) corruption in Ukraine and the idea that the country would become, at some point, a candidate for EU membership as a consequence of the agreement. Interestingly, only 23.8% of the voters found that the topic was suitable for a referendum. 3 The treaty is not yet ratified as the First Chamber still had to vote for ratification at the time of writing. For a more detailed account, see F. Lambie and S. Koole 'Reconstructie Oekraine-referendum: Hoe de nee-stem een ja wordt' online at RTLNieuws, 4/11/2016. (https://www.rtlnieuws.nl/nederland/politiek/ reconstructie-oekraine-referendum-hoe-de-nee-stem-een-ja-wordt). abstained from participation in the referendum on the Ukraine agreement. 4 Last but by no means least, the binary nature of referendum decisions results in 'winnertake-all' politics, where the views of the minority are not represented at all (and sometimes, as in the case of the Brexit referendum, this can be quite a substantial minority). In this sense, referenda also lead to much more polarization in society than parliamentary elections where provisions to take minorities and their views into account have evolved with time.
The other important question which arises after the narrow victories in important recent referenda is what to do with the (substantial) remaining minority and their views. This is a crucial issue, as in the case of the above mentioned Dutch referendum, 5 or in the Brexit referendum, 6 the results had an immediate political impact. Winning parties often claim that their victory puts all possible nuance aside. In view of such developments, Frey's plea is to bring minority views into the discussion about policy consequences after a referendum is held. He proposes that all parties involved in referenda respect minority views and enter into a discussion to seek an acceptable compromise. But does the referendum as a tool then not bite its own tail? There are pitfalls here for both leaders and voters. For leaders, we wonder how many are capable of starting serious negotiations with the 'losing' parties after winning a popular majority vote. Polarization among voters is often at a high following a referendum campaign: accommodating both 'winners' and 'losers' would seem like a 'betrayal' of the referendum promise. As we know, in the UK, Cameron could not start a debate on compromise on Brexit since he had campaigned for the yes vote, and in Italy, former Prime Minister Renzi resigned after the Italian 'no' to his proposed constitutional reforms. Not so many are willing to take the risky political path Dutch Prime Minister Rutte took in bending the Dutch 'no' into ratification.
In order to address these problems and issues, we suggest that if referenda are used on a regular basis, there should be a (possibly compulsory) pre-referendum deliberation instead of a post referendum debate as suggested by Frey. We suggest that any referenda should be preceded and always accompanied by formal deliberation (i.e., similar to citizen assembles in Ireland) ensuring that citizens are provided with both the information and the arguments that would facilitate coming to a decision. By carefully discussing the various options within a structured setting which would be part of the referendum's institutional arrangement, voters can clarify what a referendum should be about or be informed by experts about the issues or policies at stake. In addition, we suggest that only voters who have participated in the deliberation rounds, would have the right to cast a vote. In a similar vein, the proposals developed by Dryzek and his collaborators (e.g.,Dryzek 4 Others have reportedly abstained strategically in order to lower turnout below the 30% threshold, which would have made the referendum invalid. 5 Based on the total population of voters, only 19.7% of the voters voted against, 12.3% was in favor and 67.8% did not vote in the Ukraine referendum. By comparison, in the 2017 Dutch national elections, 81.9% of the voters went to the ballot box 6 In the Brexit referendum 72.2% of the voters went to the ballot box: 51.9% voted leave and 48.1% remain (or, based on the total population of voters, 37.5% voted leave, 34.7% remain, and 27.8% did not vote) and Niemeyer 2008) for experimenting with various forms of deliberation in addition to existing democratic procedures, would allow citizens to form informed opinions of local issues and problems, but also to express their emotional responses, motivations and aspirations. This is far from current practice where a limited number of signatures can trigger a call to parliament to organize a referendum over a very short period of time.
We support Frey's recommendation that referenda should be combined with deliberation focusing on potential consequences and possible compromises between the different positions and sides of a proposition. However, we realize that to accommodate citizens who may perceive themselves as losers from a specific policy can be quite hard. We believe deliberation can only help if citizens have specific value or utility based judgments on the policy in question, which are open to compromise. This is not always the case, as Dryzek and Braithwaite have shown in their analysis of the relation between discourses, political arguments and underlying values (Dryzek and Braithwaite 2000). A compromise between different positions, they suggest, is only possible on the basis of value-laden or informed judgements. If the preferred outcome is supported by a discontent-based, value-free discourse, even if others have value-based arguments, deliberation is likely to lead only to 'dogmatic reassertions of positions' (Dryzek and Braithwaite 2000, p. 262). Therefore, we can expect that in the case of majorities based on discontent rather than values, referenda might create more instability in political systems because they would contribute to further polarization.
In addition, the use of referenda as a tool of getting citizen approval for specific policies, may, in fact, lead to an exacerbation of a cleavage that is becomingincreasingly problematic in present day democracies: the gap between educated and less educated citizens. Analyses of several recent electoral outcomes that have involved politicians with a populist bent (such as Donald Trump in the US or the Brexit vote in the UK) show clear divisions between citizens based on educational status. Education influences not only how people vote, but, perhaps more fundamentally, access to education also influences income and the health of citizens. The question is whether increased use of referenda, without prereferendum deliberation on the proposed (policy) decisions, would lead to decisions exacerbating this cleavage.

Randomized Decision-Making
A third route proposed in Frey's paper is to introduce random procedures in decision-making and voter representation. Especially in the case of representation, the procedure of election by lot, would give, in Frey's view, every citizen an equal probability of becoming a member of parliament, a minister or president.
This would be a solution possibly more valuable in theory than in practice due to limitations randomly selected citizens may face in terms of competences and commitment. In terms of competences, checks-and-balances would still be in place, limiting the power of randomly selected officials, and requiring them to seek compromise and make deals with others. Running for public office may in itself help candidates acquire these vital skills. In terms of commitment, many people may not be interested in serving as a representative or government member, given the tremendous effort and time investment required. Local parties in the Netherlands, for example, face great difficulties in finding candidates that are willing to combine the rather time-consuming membership of municipal councils with their daily affairs.
Nevertheless, randomized decision-making seems an elegant way of closing the persistent gap between elites and the general public. If elites are not a fixed and defined group but can be constituted by random procedure, then citizens would presumably not feel that a privileged group they do not belong to has a disproportionate role in determining the course of politics and society. Having ordinary citizens participate in the political process based on random selection may strengthen the idea among the general public that decisions are not made by 'them' but by 'us'.
Yet we believe that the introduction of random draws, especially with regard to voter representation, will not strengthen democracy. First, it is not clear whether citizens would identify themselves to a greater extent with randomly selected citizens than with elected representatives. Research shows that in many countries the identification of voters with established political parties is very limited (Dalton 2013). It is rather unlikely that voters would identify with randomly chosen leaders either.
Second, and more importantly, in any democracy, the principles of input, deliberative and procedural legitimacy are important. It is crucial that all citizens share the belief that they can participate in democratic deliberation, are allowed to decide or vote, and recognize that outcomes are selected based on a set of commonly understood and clear rules. Based on these principles, citizens are willing to accept democratic decision-making, even if they do not agree with it.
Random procedures violate some of these fundamental principles. Having decision makers selected by the luck of the draw can be perceived by citizens who were not selected as formal exclusion from the political process.
This solution would, in our view, increase existing cleavages between groups in society and potentially also between elites (in the case of random selection those who have decided on an issue) and the rest.
The dissatisfaction of those who are not selected to public office may increase even further, eroding the common understanding on what democratic rules are about. In addition, selection by lot could lead to further alienation from the political process because citizens would realize that they have no influence on the outcome. This path leads us back to Riker's dilemma about the relative stability of institutions compared to political decisions (Riker 1980). If people start to doubt the fairness of democratic rules, it will be near impossible to maintain democracy as a model of collective decision-making. Therefore, we disagree with Frey that random procedures applied to election to political bodies would improve inclusiveness and fairness. We believe they would not have the same impact on legitimacy as the simple one-man-one-vote rule, or, more generally, the idea of equal political rights for all citizens. Making use of those procedures may even be detrimental if the openness of the democratic process to everyone is diminished.

Conclusions
The debates about the nature of the problems present-day democracies experience with citizen participation and satisfaction will undoubtedly gather further momentum and become more polarized in the coming years and decades. Frey's proposals are a welcome contribution to these debates. In our view, scholars should take a critical but independent position in this debate, as Frey does with his proposals. We do believe, however, that it is almost impossible to propose good solutions without engaging the different arguments and empirical findings and exploring the substantive as well as procedural causes of democratic discontent.
In addition to arguing for a closer link of the proposed solutions to the actual causes of citizen dissatisfaction, we propose some modifications and critical notes to some of Frey's specific proposals. In particular, we disagree with Frey on the use of referenda and random representation in collective decision-making. Concerning referenda, our critique focuses on the polarizing effects of referenda and our expectation that they have the potential to exacerbate cleavages between minorities and majorities and between educated and less educated voters. Furthermore, they may lead to unsustainable public policies and outcomes that might increase voter dissatisfaction even further. Frey's suggestion for post referendum deliberation on the decided policy goes some way to amend the problem, but in our view deliberation can only work as a pre-decision procedure and not a post referendum afterthought. The debates about Brexit which took place after the Brexit vote seem to illustrate the futility and polarizing nature of post referendum discussions.
With respect to random representation, we argue that random selection will inevitably make non-selected citizens feel excluded from the political process. Even if they have on average the same likelihood to be selected and provide input to the political process, this will be much harder to understand and explain than the simple one-man-one-vote concept. In this way, random procedures will not be perceived as satisfying the basic principles of equality, and therefore may increase perceived inequality. As democracy is also about what citizens consider to be the 'right' rules and procedures, randomization may erode citizens' recognition of legitimate decision-making.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.