1. Introduction – Market Mechanisms in Voting
A common trend in contemporary political economy involves applying market mechanisms to political institutions in an effort to improve the efficiency or effectiveness of said institutions. The motivation for this strategy arises from the notable success of markets for creating incentives and facilitating mutually beneficial exchange in conditions of sufficient information and competition. If similar conditions and mechanisms can be replicated in other domains, then perhaps the incentives and outcomes of those domains can be improved as well. The most common form of proposed implementation involves assigning property rights and creating a framework for market exchange within the relevant domain.
Until recently, the notion of applying market mechanisms to the institution of voting was unthinkable. Indeed, voting was considered the paradigm case that illustrated a domain where market mechanisms had no place (Satz Reference Satz2010: 102). Notable critic of the use of market logic, Michael Sandel, famously stipulated ‘no one defends the out-right purchase and sale of votes’ (Sandel Reference Sandel2000: 214–215). Subsequently, there have been a series of defences for vote-selling by Jason Brennan (Reference Brennan2012: 135–160), Christopher Freiman (Reference Freiman2014) and James Stacey Taylor (Reference Taylor2017 b).Footnote 1 This has broadened the philosophical debate on the use of market mechanisms in democratic institutions and fuelled vigorous responses from those critical of the prospect of vote trading. Two of the most common objections to the use of market mechanisms in voting are the equality objection and the inalienability objection. According to the first objection, the poor become more politically disadvantaged when vote trading is permitted (Archer et al. Reference Archer, Engelen and Ivanković2019). The second objection asserts that there is something intrinsically or functionally problematic with the action of selling one’s vote (Radin Reference Radin1987; Maloberti Reference Maloberti2019).
The philosophical debate on vote trading has largely proceeded independently of the public choice literature. Yet, the kinds of mechanisms in voting that public choice theorists advocate are considerably different than a simple ‘money for votes’ type of exchange. Instead, public choice theorists have moved in the direction of advocating the use of fungible vote credits that mimic some of the properties of money in order to facilitate political exchange. Notable examples include Quadratic Voting (Lalley and Weyl Reference Lalley and Weyl2018), Storable Voting (Casella Reference Casella2005), and Multi-stage Voting (Suggs Reference Suggs2020). This is relevant for the philosophical debate since the use of vote credits has markedly different properties than simple vote selling. Still, these vote credit systems were proposed as an improvement on the intended core function of vote purchasing systems, that is, to facilitate political exchange.
I begin by connecting the philosophical literature to public choice theory and demonstrate that the common philosophical objections to vote trading do not apply to vote credit systems. Consequently, if vote credit systems pose no special philosophical concerns and are more efficient than traditional voting, then they appear to offer an overall improvement in voting design. Yet, vote credit systems might falter on different grounds. I raise an alternative concern regarding the use of vote credit systems that draws on the information-generating power of voter’s impartial moral judgements. To do this I consider an example referendum that illustrates how vote credit systems change the prevalence of certain types of voting, namely, they increase interested voting at the expense of disinterested voting. I then argue that reducing the prevalence of disinterested voting could generate significant informational loss and thus poses serious negative outcomes for political choice. This includes an exploration of the importance of impartial moral judgements in democratic institutions. I conclude with an examination of the trade-offs of different voting systems and offer reflections on future proposals for voting mechanism design.
2. Two Philosophical Objections to Vote Trading
2.1. Vote Purchasing
A market in political votes is generally conceived in the philosophical literature as one in which someone may legally sell their vote to a buyer (Taylor Reference Taylor2018: 224). The buyer gains a vote and the seller gains money. That is, political influence is exchanged for an economic resource. Christopher Freiman (Reference Freiman2014) defends vote purchasing on several grounds. First, he argues the straightforward economic case that the transaction leaves both buyers and sellers better off. Second, he says there ought to be a ‘presumption of voter liberty’ where citizens can use their votes in a manner they see fit (Freiman Reference Freiman2014: 762). Thus, absent powerful reasons against vote purchasing, they should not endure a general prohibition.Footnote 2
Describing a market in votes of this sort, where a political good is exchanged for an economic good, is prominent in the philosophical literature. Yet, those defending vote purchasing tend to offer tenuous and conditional endorsements of the practice. Brennan (Reference Brennan2012: 135) argues that vote selling is not inherently wrong, as long as in doing so one ‘vote[s] for candidates or policies they justifiedly expect to promote the common good; otherwise, they must abstain’. This is a rather modest defence of vote selling, since it imposes a stringent obligation on the seller, much like expecting the blackjack dealer or the bartender to only offer their services in cases where it will clearly benefit their patrons. Similarly, Taylor (Reference Taylor2017 b) defends vote selling conditioned on the use of extensive regulations including spending caps, indexing the sale price to the wealth of the sellers, and the presence of vote-buying competition. The presence of these conditions is intended to combat what he sees as the biggest moral challenge for vote selling; the ability of the rich to buy elections and thereby politically dominate the poor (Taylor Reference Taylor2017 b: 234–236). This brings us to the first argument against vote selling.
2.2. Objection 1: The Equality Argument
The equality argument against voting markets takes many forms. Footnote 3 The most recent and articulate form is the ‘tyranny of wealth argument’ (Taylor Reference Taylor2017 b; Archer et al. Reference Archer, Engelen and Ivanković2019). According to this argument a market in voting enables wealthier individuals to gain political influence at the expense of poorer individuals (Taylor Reference Taylor2017 b: 314). Essentially, since poorer individuals have a stronger pressure or need to sell their votes, they will sell their votes for money (Archer and Wilson Reference Archer and Wilson2014: 3). By contrast, the wealthy have less need for money and can purchase a disproportionate share of votes, thereby significantly increasing their political influence (Taylor Reference Taylor2017 b: 314). The consequence is that wealthier individuals will have a disproportionate electoral influence compared with poorer individuals. We can describe this shift in political influence as having two negative implications. First, the political voice of poor individuals becomes significantly depressed because their interests, desires, preferences or judgements are less expressed in the voting booth. Second, the political outcomes of elections will disproportionately reflect the interests and preferences of the wealthy at the expense of comparatively poorer income classes. The outcomes of elections will come to primarily reflect the interests of those willing and able to pay for large numbers of votes. Hence the legal, political and economic rules will be increasingly shaped in ways which primarily benefit the wealthy at the expense of comparatively poorer income groups.
There is a further problem that unequal purchasing power presents for vote purchasing proposals. Bill Gates’s preferences on a particular topic, may be no stronger than mine, but his ability to marshal votes through purchasing will be much greater than my own. Indeed, the signal he expresses as a result of his wealth could overpower the signal expressed by tens of thousands of poorer voters who lack the power to purchase such a volume of votes. The ability and willingness to pay for more votes than others do not necessarily indicate that one’s preferences on the topic are more intense than others.Footnote 4
2.3. Objection 2: The Inalienability Argument
There is a vast array of social activities for which people have obligations and entitlements, but lack property rights. For example, parents have a set of rights involving their children, but not property rights over them, that is, they cannot sell their parental rights to others. Similarly, a juror has the right to vote on someone’s guilt or innocence, but they have no right to sell their vote to others. These rights are what can be called ‘market-inalienable’, meaning they are ‘non-salable’, though in certain conditions they may be lost or transferred to others (Radin Reference Radin1987: 1853). This is because the social purpose of these institutions is not chiefly economic gain. Parental rights are granted for the purpose of child-rearing and jury obligations are bestowed for the purposes of justice. The same argument can be extended against vote purchasing. The social purpose of voting is not to economically enrich vote sellers, but rather to participate in a ‘form of collective self government’ (Walzer Reference Walzer1974: 602). Consequently, vote selling eliminates the purpose for which the institution was meant to serve, which is to allow citizens to participate, even minimally, in public governance.
Further, allowing a legal market in votes threatens another hallmark of democracy, the secret ballot. Nicolas Maloberti argues that even if tradable vote coupons allow for de jure voter secrecy, the mere possibility of exchange means that voters cannot help but reveal information about their voting preferences, ‘contained in their decisions when faced with offers to trade’ (Maloberti Reference Maloberti2019: 203). Consequently, a market in votes erodes the secrecy of the ballot, creating the dual possibility for voter intimidation as well as clientelistic exchange (Maloberti Reference Maloberti2019: 202–205). By being able to credibly reveal support, voters and politicians would increasingly engage in forms of clientelism, where the public purse would be used to bestow private benefits in exchange for verifiable electoral support (Maloberti Reference Maloberti2019: 203). Vote inalienability protects the secret ballot and thus is an essential safeguard against political clientelism.
Even amongst those sympathetic to using market mechanisms in voting, there are few who defend an unconditional system of vote selling. A market in votes, where political influence is exchanged for economic goods is strongly susceptible to both the equality and inalienability objections. While this conception of vote trading does not appear viable due to these objections, it is not the preferred method of political exchange advocated by economic theorists. Footnote 5 Next, I turn to discuss the proposed mechanisms for vote trading that are prominent in public choice theory. After describing these vote credit systems, I turn to discuss whether they are susceptible to either the equality or inalienability objections.
3. Political Exchange and Vote Credit Systems
3.1. Voting and Preferences
Thinkers in the public choice tradition, including Wicksell, Buchanan, and Tullock, contend that as a matter of fact, political agents act chiefly with regard to their own interests and thus the tools and frameworks of economics are crucial in accurately describing political choice and behaviour (Buchanan Reference Buchanan1987: 243–246). Far from trying to minimize the role that exchange plays in political decision making, Buchanan (Reference Buchanan1987) believed that political systems ought to be designed to encourage political exchange as much as possible, saying, ‘the exchange conceptualization of politics is important in the derivation of a normative theory of economic policy. Improvement in the workings of politics is measured in terms of the satisfaction of that which is desired by individuals’ (p. 246). For Buchanan (Reference Buchanan1987: 246), the motivations and goals that agents have in both markets and politics are largely the same and differ only with respect to the structural conditions under which exchange takes place.
Although Buchanan and others argue that voting enables an exchange function similar to markets, voting mechanisms lack certain beneficial features that market mechanisms possess. The most notable of these features is the ability to capture preference intensity. The prevailing democratic principle of ‘one person, one vote’, henceforth ‘1p1v’, enables only a binary signal to be sent on any given issue. Another public choice economist, Emilia Suggs (Reference Suggs2020: 434), aptly notes, ‘while this mechanism reveals how the majority of voters feel about a certain policy, it does not reveal how intensely voters’ value certain policies over others. For a given number of distinct policies in a docket, a voter may value some policies more strongly than others, be indifferent between policies, or simply not care about certain policies altogether.’ Capturing voter preference intensity would add important information to the electoral system about which policies voters prefer. The consequence is a more efficient electoral system, where efficiency means, a better correspondence between voter preferences and electoral outcomes.Footnote 6
For those unimpressed by efficiency arguments, capturing preference intensity would have other potential advantages. Public choice theorists contend that those who have more intense preferences on a particular topic are more inclined to have ‘skin in the game’, thus capturing preference intensity will better align voter incentives (Casella and Mace Reference Casella and Mace2021: 77). It is also generally claimed that voters who express greater preference intensity on a particular topic are apt to have more knowledge and information about that topic (Posner and Weyl Reference Posner and Weyl2018: 124). According to this logic, including the preferences of voters who are indifferent or unconcerned about the outcome of a particular issue will reduce the quality of the voting signal, making it more likely to result in inefficient policy outcomes, as they do not reflect the total preference aggregation of voters (Suggs Reference Suggs2020: 435). Suggs (Reference Suggs2020: 435) concludes ‘given this concern, a voting system should not only determine how the majority of people value a certain proposal, but reduce the influence of voters who do not care about the outcomes of that proposal’.
Since The Calculus of Consent (Buchanan and Tullock Reference Buchanan and Tullock1965: 270–276) vote purchasing has been contemplated by public choice theorists, but reservations surrounding monetization made log-rolling the more preferable option for political exchange. However, within the last decade, vote purchasing proposals have evolved into more sophisticated mechanisms for political exchange that capture preference intensity without trading votes for money (Casella Reference Casella2005; Lalley and Weyl Reference Lalley and Weyl2018; Suggs Reference Suggs2020). These recent advances in public choice theory can allow for effective political exchange in a way that potentially avoids the equality and inalienability concerns that afflict vote purchasing proposals.
3.2. Vote Credit Systems
Vote credit systems enable preference intensity capture but need not involve vote purchasing. Footnote 7 A vote credit system is any system where a voter uses a fungible pool of credits that can be cast across the electoral options in any way they desire. For example, a voter may receive a ballot with 5 issues under consideration and may allocate 10 vote credits in a manner of their choosing. They could spread the credits evenly by using 2 vote credits on each issue or spend them disproportionately on selected topics, favouring some issues and ignoring others altogether. The fungibility of vote credits enables voters considerably more flexibility to express their preferences. Just as a shopper chooses a basket of goods, a voter can now choose a basket of proposals. Their selections represent not only their views on particular topics, but the relative weight of importance that they place on particular topics.
All vote credit systems allow voters some discretion in allocating their vote credits, but the mechanisms and constraints for allocation vary considerably. One notable proposal (Casella Reference Casella2005: 391) is storable voting, where voters can decide to abstain on topics and store votes for future use on topics for which they have stronger preferences. As opposed to typical vote purchasing proposals ‘in a system of storable votes, votes are traded intertemporally, not interpersonally’ (Casella Reference Casella2005: 392). Another proposal, quadratic voting, was originally intended as a vote purchasing scheme, but was later adapted into a vote credit system that functions without the buying or selling of votes (Posner and Weyl Reference Posner and Weyl2015). In quadratic voting, voters can allocate vote credits in any manner they desire, but the strength of each additional vote spent within a particular topic decreases quadratically (Posner and Weyl Reference Posner and Weyl2018: 82–120). Thus, one vote on topic Q costs one vote credit, two votes on topic Q costs four vote credits, three votes on topic Q costs 9 vote credits, and so on. The motivation for quadratic voting is to facilitate exchange amongst voters but without allowing for special interest capture in a way that might occur under a vote credit system where vote credits are weighted linearly (Lalley and Weyl Reference Lalley and Weyl2018: 34). The latest innovation in vote credit systems is a two-stage multi-voting proposal (Suggs Reference Suggs2020: 6). Voters would allocate their vote credits across specific policy topics in the first stage, and then allocate their vote credits among the options within each policy topic in the second stage.
Although each proposal operates a bit differently, the fundamental logic taking place in all vote credit systems is the same; what Buchanan (Reference Buchanan1987: 246) called ‘politics as exchange’. The allocation of vote credits spread across topics allows voters to implicitly exchange influence on items that they care less about for items that they care more about. If I care a lot about topics A and C, and you care a lot about topics B and D, our vote credit allocations into those topics allow for an effective exchange. By contrast, under the 1p1v system, our ballots would not enable effective political exchange, but rather my weak preference on a particular topic would count the same as your intense preference on that topic, and vice versa. The upshot is that vote credit systems enable more efficient political exchange by including information about preference intensity. The only question remains, do they avoid the standard philosophical objections to vote trading?
3.3. Vote Credit Systems Avoid the Equality and Inalienability Objections
Any voting system that allows the buying and selling of votes faces the equality objection. However, vote credit systems do not actually exchange votes for money, they exchange relative political influence. If I sell you my vote for cash, I am exchanging my political influence for something else, maybe gas money or pocket change. Laurence and Sher (Reference Laurence and Sher2017: 196) comment on this effect: ‘the most pressing and intuitive concern with vote buying schemes … involves the way they transform economic inequality into political inequality’. However, vote credit systems don’t allow for the exchange of votes for dollars, they allow for the exchange of political influence in one area for the political influence in another area, and without the need for an explicit trade.Footnote 8 Under a pure vote credit system, there is no way for me to sell you my vote on the ‘cigarette tax referendum’ for gas money or pocket change. However, if I decide to forego my vote on the ‘cigarette tax referendum’ so I can put an additional vote on the ‘elimination of cash bail’ proposal, then I have effectively exchanged my political influence on one topic for greater influence on another topic. Since there is no possibility of selling votes, the lower income classes cannot exchange political influence for money. Thus, the political expression of the lower income classes is not reduced, nor can the wealthy outright purchase disproportionate political power. The poor can redirect their votes towards topics they care more about and thus could effectively engage in political exchange with the rich. However, given that the poor outnumber the rich, there is little argument that they are somehow at a disadvantage in this system, at least not any more than they already are in the prevailing 1p1v system. Thus, pure vote credit systems do not suffer from the equality objection.
Can the inalienability objection be levied at vote credit systems in the same way that it is levied at voting markets? In a vote credit system, the individual never buys or sells a vote, instead, amongst a number of issues, they decide where to allocate their vote credits. There is no alienation, because voters continue to express political influence and have no market incentive to opt out. Further, as there is no vote purchasing, there is no concern about maintaining the secret ballot. Far from alienating the individual from political choice, one can view vote credit systems as enhancing political expression because voters can now also express the intensity of their preferences on a particular topic. Under a vote credit system, a voter is free to allocate a single vote on each topic as they did under the 1p1v scheme, or they may decide to concentrate their voter expression into preferred topics.
Just as with the equality argument, the inalienability argument does not apply to vote credit systems and for similar reasons. Vote credit systems do not exchange political representation or political expression for some other good, resource or commodity. Instead, vote credit systems allow for voters to exchange political representation or expression amongst political topics or candidates. Vote purchasing allows for exchange between the political sphere and the economic sphere, while vote credit systems do not. Under vote credit systems, the exchange amongst voters is confined to the political sphere.
If there is a problem with vote credit systems, it is not that they exchange political influence for economic gain. Instead, there is something wrong with the prospect of conceiving political influence as an area for which we should exclusively rely on the concept of ‘exchange’. I now turn to elaborating the role that impartiality plays in institutional settings.
4. Impartiality and its Institutional Purpose
Impartiality is endemic to a well-functioning democratic society. Impartiality constitutes the moral backdrop of almost every formal institutional setting therein. In the court of law, judges are required to recuse themselves in cases that present a conflict of interest or often even where there may just be the appearance of bias. Jurors with any partial or preferential ties to the plaintiff or defendant are systematically screened out by the opposition’s lawyers. Footnote 9 Politicians are encouraged and sometimes required to put their financial holdings in trust so their policy decisions are not unduly influenced by their pecuniary interests. Publishing academic and scientific research requires the authors to state any potential conflicts of interest and undergo double or triple blind review. Even in the business world, publicly traded companies have placed strong emphasis on eliminating conflicts of interest, nepotism, or other forms of favouritism in their hiring practices. Of course the presence of these impartial screens does not always succeed, but their purpose is clear: to prevent these institutions from being co-opted by private interests, or interests for which the institution was not intended to serve.
What’s notable is that the presence of impartiality in these institutional settings serves to screen out arbitrary bias, preference, discrimination or favouritism. The purpose of being impartial or having impartial adjudicators is to prevent or screen out the presence of corrupting factors, factors that stymie or undermine the authentic purpose of the institution. This is also how impartiality functions in both Harsanyi’s and Rawls’s veil of ignorance, where arbitrary features of one’s own identity are obscured, so that one’s judgement within the original position is not distorted (Rawls Reference Rawls1999: 122; Grant et al. Reference Grant, Kajii, Polak and Safra2010: 1939–1940).
Voting, at first, appears to be different from the other aforementioned institutions. One of the primary purposes of voting is to allow citizens to express and assert their interests so that they are represented, in aggregate, by the adopted policy or elected representative. Footnote 10 Thus, voting is not a process for which we would want to screen out the presence of private interest. However, the desirable representation of private interest in the electoral system does not preclude the desirable representation of impartial disinterested citizens. Our question here is whether or not impartiality has any role to play in the electoral process.
As we have seen, public choice theorists tend to hold that the presence of unaffected individuals ‘do not contribute new information about the preferences of the majority’ (Suggs Reference Suggs2020: 2). Yet, this presumes that the only legitimate purpose for voting is to express and assert one’s own interests and preferences. In other forms of voting, such as takes place in a jury trial, impartiality is key to adding important informational content. Jurors are selected on the basis of their disinterest, so that their judgements, whether they be applied to fact, law or morality, are unobscured by the presence of preferential ties. Is the same true for voting? Does the presence of disinterested voters, those who have few or no ties to the issues at hand, contribute valuable and relevant information to the electoral system? Adam Smith (Reference Smith1982) argued that moral distance, real or hypothetical, was essential to the generation of legitimate moral judgements, remarking:
We can never survey our own sentiments and motives, we can never form any judgment concerning them; unless we remove ourselves, as it were, from our own natural station; and endeavour to view them as at a certain distance from us. (Smith Reference Smith1982: 110).
Smith was speaking about how individuals can come to morally assess their own actions. He argued that the only way to achieve this is for someone to reflectively divide themselves into two persons, the agent whose character is under scrutiny, and the spectator who examines and judges the actions of the agent, at a distance (1982: 113). Smith’s account of moral evaluation requires that the source of scrutiny come from the outside, either in the case of actual external impartial spectators, or in the case of honestly conceiving ourselves as impartial spectators and judging ourselves accordingly (Sen Reference Sen2002: 450–451). We need not hold, as Smith did, that this is the only form of moral evaluation, but we can see how impartial judgements from a distance can play a powerful role in moral evaluation. Indeed, as we’ve discussed, most formal democratic institutions operate embedded within conditions of impartiality.
Next, we turn to explore what impartial spectators, not those conceived in our own minds, but those real spectators that we call our fellow citizens, can contribute to those parts of the public domain for which they are disinterested.
5. Example Referendum: The Case of E-Cigarettes
5.1. Two Voter Types
On a given political topic or issue, I distinguish between voters using two broad categories, interested voters and disinterested voters. Interested voters possess a substantive stake in the matter at hand, while disinterested voters do not. Interested voters can further be distinguished into two types, the materially interested and the ideologically interested. For our purposes here, I consider someone to have a material interest whenever they have a strong financial, familial, or health-related interest in the issue at hand. For example, a voter working in the tobacco industry would have a material interest to vote in particular ways on tobacco-related legislation.
Ideologically interested voters also have a substantive stake in the issue at hand, but it’s not derived from their financial, familial or health interests. Instead, someone is ideologically interested when the issue at hand affects their deeply held beliefs and values. For example, someone who strongly identifies as “pro-life” or “pro-choice” is an ideologically interested voter when it comes to topics closely related to abortion. Having political opinions on a given topic is not itself sufficient to be ideologically interested on that topic. Rather, ideological interest occurs when the presence of one’s deeply held beliefs or values threatens to obscure one’s independent judgement on a particular topic or issue.
A disinterested voter is any voter who does not have a material or ideological interest at stake in the issue at hand. These voters are not necessarily moderates, nor need they be ideally situated impartial political judges. They can range across the political spectrum and they can be members who are affected by the policies; it’s simply that their material or ideological stake is comparatively low. Any given voter will likely be a disinterested voter at some point in time, depending on the issue under consideration and their relation to it. Under this framework, we can only say a voter is materially or ideologically interested in relation to some topic or issue; this is not a stand-alone category.
Voters may not always be described as purely interested or purely disinterested. For some topics, one’s level of interest lies on a spectrum. Suppose a tariff on avocados increases their price by 2%. This small effect is unlikely to make an interested voter out of the average grocery shopper. By contrast, a bill that would affect an individual’s eligibility for health insurance, constitutes a substantive stake in a person’s interest and plausibly qualifies them as an interested voter. Yet, there are cases between these examples which are more ambiguous. Accordingly, this piece does not suppose that voters always fall cleanly into one category or the other, but may lie on a spectrum between the two extremes. The upshot is that the less materially or ideologically interested a voter is, the more apt they are to vote on the basis of a wider range of considerations.
As I use the pairwise terms, ‘interest’ and ‘disinterest’, I am referring only to whether or not voters have a substantive stake in the topic, ideological or material, and not whether they are psychologically attentive to political or civic issues. Thus, disinterested voters can be civically engaged and informed, but have no vested interest. For example, they may be politically attentive but undecided, or merely possess slight, but non-intense preferences. Finally, the presence of an interest for a voter is not deterministic, but it is suggestive. Voters can and do vote contrary to their material or ideological interest, but they are, ceteris paribus, more likely to vote for their material or ideological interest than against it.
5.2. Voter Information and Judgements on E-Cigarettes
Let’s begin by considering a hypothetical state referendum where one of the propositions under consideration is as follows: ‘the state government implements a tax on electronic cigarettes and dedicates the subsequent revenue to health programmes’. Footnote 11 Materially interested parties against the proposition include individual buyers and commercial sellers of e-cigarettes. Materially interested parties in favour of the proposition would include parents of teenage children who wish to discourage smoking, as well as healthcare organizations that receive revenue from the new law. Ideologically interested parties would include those possessing a deeply held belief in economic freedom such as libertarians and the local chamber of commerce. Ideological proponents of the new law would include healthcare providers with strong negative attitudes towards smoking as well as other health advocacy groups.
The information and motivations of interested voters are highly concentrated. To illustrate, the frequent smoker knows the joys (and harms) of his daily experience of smoking; he may on balance enjoy smoking. He also knows how difficult quitting can be and has likely considered that an increase in price is unlikely to discourage him, but will simply cost him more money. Similarly, a doctor at a cancer ward is acutely aware of the long-term medical risks of smoking, he observes them every day. For the smoker and the doctor, these are very likely to be the considerations that motivate their respective and opposing votes. They each have large amounts of relevant information, but for each of them the information is highly concentrated to a narrow band of considerations.
What about disinterested voters, who are they and how will they vote? Disinterested voters include everyone else who we did not mention, that is, the vast majority of the voting population. The opinions of these voters are likely to be highly variable. They may not care about the issue one way or the other, or they may have strong opinions about the issue, despite not having a material or ideological interest in it. The point is that we don’t have a strong prima facie reason to know how these voters are going to behave, and that is precisely because they are disinterested. But given that these voters largely exist on the periphery of these issues, what information can they possibly add?
A disinterested voter likely knows a bit about smoking. They may have smoked in the past, had a family member who smoked, or may even occasionally enjoy a smoke at the bar from time to time. They likely know that smoking comes with some health costs and they may know that some smokers love smoking while others regret having the habit. Some disinterested voters probably know someone who has replaced a tobacco habit with a less harmful electronic cigarette habit. Inversely, other disinterested voters may know someone who was lured into smoking by the false perception that electronic cigarettes pose no health risks.
Further, disinterested voters, or comparatively less-interested voters, are also likely to have access to some kinds of information that interested voter’s lack. Smokers are unbothered by second-hand smoke, (or second-hand vapour in the case of e-cigarettes), but these may present a slight nuisance, or mild enjoyment, for non-smokers.Footnote 12 These preferences are typically too small to make these voters interested, as the smell does not constitute a substantive stake for these voters.Footnote 13 Yet, although these preferences on second-hand smoke tend to be small and diffuse, nearly everyone has some preference with respect to it. This information could play a role in how disinterested non-smokers vote. This is illustrative of how disinterested voters can contribute information about small but widespread harms or benefits to the electoral calculus.
In summation, we may make some tentative observations. The lack of substantive interests on the part of disinterested voters means that they have less of an incentive to gather information on the subject, but it also means that they are less likely to confine their vote to a narrow set of considerations, since they are not motivated by material or ideological interest therein. Thus, the opinions of disinterested voters are plausibly more apt to include a broad array of considerations. By contrast, interested voters have more reason to vote on the basis of their concentrated interest and to pay less attention to other considerations. If these observations hold, the presence of interested voters adds informational depth while the presence of disinterested voters adds informational variability. A reduction in either type of voting results in a particular kind of information loss.
5.3. Voter Allocation on E-Cigarettes
Recall our state referenda. Imagine that the given docket contained ten propositions, where the electronic cigarette proposition was just one of the issues under consideration. Suppose that instead of a 1p1v system, a VC system was in place such that voters could allocate vote credits across the docket. Each voter is given ten vote credits and they can allocate these credits across the ten propositions in any manner of their choosing. Voters could decide to vote just once on each topic, dedicate all ten credits to a single vote, or any allocation in between these extremes. Assume that each vote credit is weighted the same, regardless of how many vote credits that someone allocates to a proposition.
What would the consequence of the vote credit system be for the electronic cigarette proposition? It would of course depend on what other issues were on the ballot. Each voter would have to decide for themselves how they valued the relative importance of each of the ten propositions, including the e-cigarette tax proposition. Those with a strong material or ideological interest on the e-cigarette tax proposition are the kinds of voters who are more likely to allocate votes to this proposition compared with other ballot items. These would include all of those groups previously mentioned: smokers, businesses with relevant commercial interests, ideological groups dedicated to commercial freedom, ideological groups that oppose smoking, health advocacy groups, healthcare professionals, and the parents of teenagers who smoke or might smoke.
What about disinterested voters, are they more or less likely to vote on the e-cigarette tax proposition? If a voter is considering whether to spend her vote credits on the e-cigarette topic, of which she may have an opinion but no stake, or a topic relating directly to her material or ideological interest, where she has both a stake and an opinion, where is she more likely to vote? All things considered, she seems less likely to vote on the e-cigarette proposition since she could otherwise direct her vote credits to a proposition where she has a greater material or ideological interest. A ballot with nine other measures is likely to have some other topics that bear more directly, or more substantively, on her material or ideological interest than this one. Thus, she can substitute her vote credit on a topic for which she is disinterested in exchange for an additional credit on a topic for which she cares immensely.Footnote 14 Indeed, this is the motivation that public choice economists have for implementing a vote credit system (Posner and Weyl Reference Posner and Weyl2018: 105–108; Suggs Reference Suggs2020: 435).
The upshot is that under a VC system voters give up a say for something they care about less in exchange for something they care about more. The consequence of implementing a VC system is that for every topic, the proportion of interested voting increases relative to disinterested voting. Thus, disinterested voting becomes considerably less electorally significant than interested voting under the VC system.
5.4. The Social Implications of the VC System for E-Cigarettes
What difference should we expect on the e-cigarette vote when the VC approach is applied instead of the 1p1v approach? Due to the new incentives, we can reasonably expect that interested voters will play a comparatively larger role in the electoral outcome and the effects of disinterested voting will be comparably reduced. Given that the judgements of disinterested voters play less of a role in determining the electoral outcome under the VC model, we can get a general idea of what forms of social signalling will be diminished or lost.
Diffuse considerations related to the advantages or disadvantages of an e-cigarette tax will play a less significant role. The proportion of voter expression based on considerations related to second-hand smoke/vapour, or the public nuisance of smoking, will likely be diminished. Other periphery judgements will also be diminished, such as any perceived benefit that smoking brings to the social milieu or cultural character of the city. Further, voter expression that involves balancing or weighing the general interest will also probably be reduced. For example, votes cast on the basis of judging that the much-needed revenues are more important than the inconvenience of the tax, would play less of a role. Similarly, the judgements of those determining the opposite would also be diminished. The upshot is that there is less total information being provided from voters who vote on a broader array of considerations than those who vote narrowly on the basis of a direct material or ideological gain.
Yet, the information loss in the election is likely not the most significant difference between the 1p1v and VC approaches. The largest difference will likely be the legislation that is under consideration in the first place. As the disinterested proportion of the voting base decreases relative to the interested proportion, we should expect to see a change in the calculus of policymakers. Legislation is typically crafted by an interested group or alliance of interested groups and relevant partisans. Under a typical 1p1v situation, the legislation has to be written in such a way that enough disinterested voters will vote in its favour, thus the crafters often have to moderate its content. But under a VC situation, the calculus changes. As disinterested voting decreases relative to interested voting, the legislation does not need to capture more disinterested voting, it needs to capture more interested voting. Thus, the general majority of voters becomes less relevant as a deciding faction. Instead, a faction called the interested majority becomes more relevant. By interested majority, I am referring to that majority coalition within the group of interested voters. Thus, as the relative strength of interested voting increases, the coalition in power will write legislation that is increasingly attentive to the advantage of the interested majority and less concerned about the will of the general majority.Footnote 15
As this logic applies to our e-cigarette example, if the coalition in power is sympathetic to the healthcare lobby, its affiliates and related interest groups, then the proposed legislation will be written primarily to motivate that interested majority. For example, the tax may be particularly steep or the regulations governing e-cigarettes and smoking may be particularly onerous and draconian. By contrast, if the coalition in power is sympathetic to smokers and smoking interests, then the proposed legislation will be extreme in the other direction. The subsequent legislation will likely vastly underestimate the social costs of smoking and be extremely favourable to smoking industry groups to the detriment of public health. Under the VC model, the reduced presence of disinterested voting means that the interested majority has a much weaker incentive to moderate its legislation, making the resulting policies more extreme and one-sided.
Our hypothetical e-cigarette referendum has illustrated the mechanics of the VC model as well as some of its implications. I now turn to generalize these features and discuss them in more abstract terms.
6. Three Social Functions of Disinterested Voting
The presence of disinterested voting could affect political choice in many ways, but there are at least three functions that disinterested voting can perform: a collective action function, a social inclusion function and a moderating function. I explore each of these functions in turn.
6.1. The Collective Action Function
One type of disinterested judgement that people tend to make in the context of political topics are those judgements that appeal to the general welfare. Of course, what constitutes the general welfare is open to reasonable disagreement. Nevertheless, the general form that these arguments take is quite familiar. Negatively, they involve limiting actions or behaviours that are privately beneficial, but constitute public threats or harms. These plausibly include policies that curtail crime, wildfires, pollution, global warming, second-hand smoke or other negative externalities. Positively, they involve generating systems that promote actions or behaviours whose private incentive is small, but whose public benefit is large. These include policies to create public goods or other positive externalities such as national security, knowledge, disaster prevention or infrastructure.
In most cases, limiting a particular negative externality will be to one particular group’s concentrated disadvantage and to the larger group’s diffuse benefit. Under the VC model, there is an increase in interested voting relative to disinterested voting and thus we would expect an increase in the adoption of policies that create concentrated benefits for interested minorities but with diffuse costs for disinterested (or comparatively less-interested) majorities. Similarly, positive externalities and public goods will find decreased support under the VC model, especially in circumstances where the source of funding is concentrated to this or that group. The upshot is that disinterested voters in the 1p1v approach play a collective action function. Disinterested voters have less of an incentive to ignore the diffuse public benefit than interested voters do. Measured absolutely, disinterested voters likely perform the collective action function sub-optimally, however by comparison they probably perform it considerably better than interested voters. As the power of concentrated interests grow relative to the power of the disinterested majority, less political weight is given to diffuse social costs and more weight is given where benefits are concentrated. Consequently, the problem of regulatory capture via political rent-seeking is likely to worsen under the VC model.
6.2. The Social Inclusion Function
Legislation voted on by citizens often has tremendous implications for non-voting stakeholders. Children, immigrants, foreigners, animals, the unborn, the incarcerated, future generations, ancestors, even God, can all be regarded as having a stake in political outcomes voted on by the citizens of any particular polity, despite not having formal representation. The topics that are under consideration related to these groups often deal with their rights and deeply held interests, such as whether criminal defendants have to give cash bail, whether parolee’s receive voting rights, or whether undocumented immigrants can have access to social services. Footnote 16 Is political exchange the correct mechanism by which issues of deep moral valence get decided? Public choice theorists may argue that these situations will get decided by political exchange no matter what, it just might as well be efficient exchange. Therein lies the problem. Non-voting stakeholders cannot vote, and therefore they cannot benefit from political exchange as they lack the relevant currency. What are the implications of moving from a 1p1v model to a VC model for non-voting stakeholders?
The first implication is that these issues regarding social inclusion will get decided with the wrong kind of information. As interested voting increases relative to disinterested voting, voter expression will reflect voter’s personal preferences more than their ‘at-a-distance’ moral judgements. If voting on topics of social inclusion should be more analogous to the kind of reflection that takes place in a jury than the kind of decisions that take place in a supermarket, then the VC model moves us in the wrong direction.
The second implication is that non-voting stakeholders have a decreased capacity for moral appeal. Given that non-voting stakeholders lack formal representation, they need to rely on other voters for representation.Footnote 17 Would an increase in interested voting and a decrease in disinterested voting result in greater or lesser consideration of the rights and interests of non-voting stakeholders? As interested voting increases in proportion, non-voting stakeholders would only stand to benefit in situations where their interests incidentally align with those of voting stakeholders. However, in any circumstance where the interests of non-voting stakeholders conflicted with interested voters, they would likely find themselves at a severe disadvantage. In a 1p1v system, non-voting stakeholders have a potential recourse as they can appeal to voting non-stakeholders, i.e. disinterested voters. Since disinterested voters are not beholden to a particular interest in the case at hand, they make attractive targets for moral appeal. Non-voting stakeholders or their representatives can appeal to the moral judgements of disinterested voters. When disinterested voters consider the interests of non-voting stakeholders we can say that they are performing an inclusionary function. By contrast, interested voters are less attractive targets of a moral appeal, especially where they have an incentive to ignore the interests of non-voting stakeholders. Their interest is a potentially biasing factor in their moral judgement. The relative reduction of disinterested voting would make it harder for non-voting stakeholders to make such an ethical appeal, and thus would be more subject to the whims of interested voters.
There is an important qualification to be made regarding the social inclusion function. Ideological interest groups are often formed specifically to protect the interests of certain non-voting stakeholders. Consider the pro-life or animal rights voting coalitions. The former purports to vote with the interests of unborn babies and the latter purports to vote for the interests of animals. Consequently, there is an argument to be made that a relative increase in interested voting could potentially increase protections for non-voting stakeholders. This is plausible for groups that have strong ideologically interested coalitions. But there are also problems with this line of response. First, ideologically interested voters may perform a social exclusionary role as well as an inclusive one. Consider for example, coalitions of populist or nativist voters that vote on the basis of excluding the rights or interests of non-voting stakeholders, such as for immigrant populations. Second, there are many non-voting stakeholders whose interests often lack strong ideological support; the incarcerated, immigrants, foreigners and future generations are notable examples.
A move from 1p1v to VC systems would mean that there would be a reduction in disinterested voting and hence less of a target audience for independent moral appeal by non-voting stakeholders. This does not mean that disinterested voters may always be relied upon to protect the interests of non-voting stakeholders, as they themselves might be subject to other biases or mere ignorance of the plight of non-voting stakeholders. Still, the point remains that because of their lack of motivated interest in the topics at hand, disinterested voters are comparatively better targets for moral appeal for non-voting stakeholders than interested voters. It is difficult to ascertain whether the presence of ideological interest groups would be sufficient to safeguard the interests of non-voting stakeholders in a VC system. Prima facie, it would seem that non-voting stakeholders that already have the support of an ideological interest group might do better under VC systems, while those lacking such support would fare worse.
6.3. Moderating Function
The social functions that disinterested voters perform apply not only at the moment of political choice, but also dynamically affect the kinds of propositions that are even brought to the fore. The manner in which referenda arrive on the ballot is itself a process involving considerable political calculation. Those interested parties, groups or legislators drafting proposals have to write them in such a way that balances two aims: first, that the proposal satisfies their coalition’s goals, both material and ideological, and second, that the proposal has a reasonable chance of actually being selected by the majority of voters.
Consider the difference between writing a proposal that needs to obtain the majority of votes in a 1p1v system as opposed to obtaining that majority under the VC model. In the former instance, the drafters know that a larger number of people will be voting on their proposal and that this group of people is likely to be considerably diverse when it comes to their interests, values, goals, judgements and considerations. Thus, the drafters know that their policy cannot be so narrowly written as only to motivate their own interest group, it must also be acceptable to a sufficient number of disinterested voters, who are voting on a wider set of reasons and considerations. By contrast, when drafters are working under a VC model, their calculus is considerably different. Knowing that the amount of disinterested voting is considerably lessened, the result is that a much narrower group of people is likely to vote on the topic, those having a strong material or ideological interest at stake. Hence, the drafters need to write the legislation in such a way that highly motivates those of the interested majority to vote for their proposition. The more that the legislation embodies the interested majority’s political aims, the greater incentive that voting bloc has to allocate more vote credits to that particular piece of legislation. Further, as the relative proportion of disinterested voters diminishes, the interested minority has less of a target group to which it can appeal or lobby on its behalf. Hence, the drafters have an incentive to write the proposed legislation in such a way that is sufficiently motivating for a majority of the interested parties, but in a way where little or no attention needs to be paid to interested minorities, nor to the diffuse matters concerning the general population, or considerations relevant to non-voting stakeholders.
Some may be unimpressed by the idea of moderating political outcomes by confusing moderation with centrist politics. The moderating influence is important, not because it gets us closer to the ‘political centre,’ but because it’s more apt to result in policies that include considerations from a broader array of interests, including those of the interested minority. Since under a VC system interested minorities have less of a capacity to appeal to disinterested voters, the drafters in the interested majority have less of an incentive to write the policy in a way that is sensitive to considerations of the interested minority, as well as that of a wider constituency. Consider for example, what the consequences of a VC model would be on polarized topics in regional jurisdictions where interested majorities required less political buy-in from the general voting populace. Under the VC model, proposed legislation related to polarized issues such as firearms legislation, abortion, drug legalization, criminal justice reform and LGBTQ issues, would likely be more extreme in either direction.
There is likely a natural limit on how one-sided the interested majority could make legislation under the VC model as severe extremification may transform previously disinterested voters into interested voters. Still, under the VC approach, extremification faces less resistance because comparatively disinterested voters incur a cost for voting on these topics, namely they give up votes on topics for which they are more interested. Therefore, policy proposals must become so extreme that previously disinterested voters become interested enough to allocate vote credits to these issues.
By contrast, under a 1p1v system, the presence of a larger group of disinterested voters enhances their capacity to play a swing-voter function. If the interested majority advances an agenda that is too extreme or too disadvantageous to the other side, the interested minority has a larger pool of disinterested voters to which they can appeal. Unlike voters in a VC system, voters in a 1p1v system, incur no opportunity cost when voting on a topic for which they are disinterested (because they only get 1 vote on every topic anyway). The possibility of this appeal plays into the political calculations made by the drafters in the interested majority. Proposed legislation needs to be written in such a way that disinterested voters are less likely to be swayed by the appeal of the interested minority. Thus, the mere presence of disinterested voters moderates proposed legislation. It is in this way that 1p1v actually moves political choice closer to what public choice theorists (Buchanan and Tullock Reference Buchanan and Tullock1965: 276) generally regard as the optimal point, unanimity.
7. The Trade-offs Between Interested and Disinterested Voting
VC systems are more likely than 1p1v systems to incur information loss related to the beneficial social functions simply because disinterested voters are less biased than interested voters. That is, the fact that a voter is disinterested on a topic, means that they have no partial interests that stymie their evaluation of other relevant considerations. Yet, the increase in interested voting under VC systems is a feature, not a bug, and comes with certain purported advantages. VC systems are designed to facilitate effective political exchange between voters such that each voter can obtain greater influence on topics they value more in exchange for giving up influence on topics they value less. Three justifications are offered for moving to this approach: the efficiency justification, the information justification and the incentive justification. I discuss each of these justifications in light of the difference between interested and disinterested voters.
According to the efficiency justification, VC systems improve alignment between voter preferences and electoral outcomes. As mentioned earlier, efficiency in this context does not refer to output or utility, but only to an increased correspondence between what voters prefer and whether those political outcomes were selected. VC is purportedly more efficient because unlike 1p1v it not only includes voter’s preferences, but also the intensity for which voters have those preferences.
Yet, there is a paradox lurking in the public choice framework. Those who have the greatest intensity of preferences on a particular issue are those who have a strong vested interest in that particular issue. For example, tobacco companies have the most to gain or lose when it comes to tobacco legislation. Is this a reason for or against, giving those who work in the tobacco industry more or less of a say in tobacco legislation? Intuitively, we should say those in the tobacco industry should have less of a say on the legislation, since they are the most interested and hence the most apt to corrupt the legislation. Yet, according to the VC model, those in the tobacco industry would be able to maximize their say on tobacco legislation in exchange for giving up influence on other topics. Is the VC model simply a way of streamlining rent-seeking, but on a fair and open playing field? The difference between the rent-seeking in the VC model and its traditional manifestations is the level of formalism and transparency. Where traditional rent seeking and regulatory capture is confined to backroom political dealings obscured from public view, the VC model can be viewed as democratizing rent-seeking, such that everyone can participate in trying to maximize their influence on topics for which they have the greatest interest. This transparency comes with a natural safeguard; when the social costs of a policy are on the ballot, they are more transparent, and voters who stand to lose can directly vote against them. However, the problem is that social costs which are large but diffuse are less likely to take precedence in the minds of voters, and thus we can reasonably expect that the average electoral outcome will accompany an increase in externalized social costs. Another problem with the efficiency justification is that it assumes that preferences are the only important form of information and ignores the value of voters’ disinterested judgements.
The other two justifications for implementing a VC model, the information and incentive justifications, have to do with the constitutive difference between interested and disinterested voters. Of course, the voters themselves are all the same, but their knowledge and motivation differs depending on the topics for which they vote. As we move from 1p1v to VC, there is an increase in interested voting and a decrease in disinterested voting. When people vote on their interest, they are thought to have more information because they care more, and they are also thought to have better incentives, because they have more skin in the game.
Since interested voters have a greater stake in the topic at hand, it is quite plausible to suggest that they are more motivated to seek out relevant information. Yet, upon further investigation, it becomes less clear that this is actually an advantage. First, the nature of their interest likely affects the kind of information they seek out, meaning that they focus on information relevant to their interest to the exclusion of information that is not relevant to their interest, or that is contrary to it. Second, given their interestedness in the topic, interested voters are likely more susceptible to engage in motivated reasoning and be subject to confirmation bias. Disinterested voters have less of an incentive to gather information, but because they’re less concerned with finding a particular result, their disinterest makes them less susceptible to engage in motivated reasoning. Thus, while disinterested voters are likely to be more uninformed than interested voters, interested voters are more apt to be misinformed. According to a recent study, there is evidence to suggest that ‘uninformed individuals are more likely to update their beliefs than misinformed individuals after exposure to corrective information’ (Li and Wagner Reference Li and Wagner2020: 646). Finally, there is considerable social scientific evidence of powerful informational advantages that accrue from the presence of cognitive diversity (Landemore Reference Landemore2012: 160–166). Given that interested voters will tend to cluster around the same types of interests and experiences, and disinterested voters are by definition more diverse in this respect, we should expect greater informational advantages from cognitive diversity to occur in situations where there is more disinterested voting. Taking all of these considerations together, the total informational advantage may belong to disinterested voters.
Second, interested voters are more likely to have skin in the game, but here again the question becomes whether this makes them better or worse voting agents. The wolf has a lot to gain when the sheep gate is left open. Yet, that does not mean that giving the wolf greater controlling vote over the sheep gate results in desirable outcomes. This is especially true if the sheep themselves cannot vote, as in the case of non-voting stakeholders. Does an increase in interested voting better align the incentives of political choice or result in greater political corruption? This depends on whether the issue at hand comes with the possibility of cost-externalization. As the possibility for cost-externalization falls, the desirability of interested voting increases as political exchange is more likely to be pareto-optimal. Footnote 18
Whether the 1p1v or VC model results in better political outcomes is likely contingent on each situation of political choice. In situations where political outcomes have little potential for diffuse costs, where all stakeholders can vote, and where there is roughly equal representation of interested coalitions, then political outcomes are likely improved by the VC model. The further one moves from these conditions, the more likely that the VC model will have deleterious social outcomes compared with the 1p1v model. Additionally the effectiveness of disinterested voters also depends on other contextual features, such as whether voting is mandatory or voluntary, and the level of information that disinterested voters possess. Disinterested voting is likely more desirable in situations where voting is voluntary and where voters generally possess a high amount of information and less desirable where voting is mandatory and voters possess low levels of information.Footnote 19
8. Conclusion: Political Exchange, Impartiality and the Future of Democracy
Traditional objections to vote trading stemmed from concerns about inequality and inalienability. Proper mechanism design in vote credit systems can ameliorate these concerns by confining exchange to the political sphere. However, any formal system of political exchange cannot avoid increasing interested voting at the expense of disinterested voting. The very purpose of exchange requires that parties give up something they prefer less to something they prefer more. Hence, when private interests come to dominate political choice, judgements based on disinterest and deliberation at a distance are diminished. Consequently, to the extent that disinterested voters exhibit beneficial social functions related to collective action, social inclusion or legislative moderation, these functions would be diminished under a VC system.
Absent empirical verification, how confident can we be in these conclusions? This piece borrows the same assumption that predominates legal theory, which is that the judgements of interested parties are more subject to bias and partiality than evaluations by disinterested citizens. To the extent that we can be confident in this assumption, is the extent to which we can be confident about the general contours of the conclusions drawn herein. Namely, that VC systems encourage interested voting at the expense of disinterested voting and this diminishes the potential beneficial effects that come with disinterested voters’ impartial judgements.
Nevertheless, there are potential confounding factors that might obscure or mitigate these general results. Simply because disinterested voters are comparatively less biased than interested voters does not mean that the beneficial social functions of disinterested voting always shine through. Perhaps disinterested voters are so uninformed (Somin Reference Somin2013) or unengaged that these social functions generally play a negligible role in political outcomes. That is, although disinterested voters are better positioned to perform these social functions than interested voters, does not entail that they always will perform these functions. Consequently, the extent to which disinterested voters exhibit these social functions requires empirical and experimental examination from which this piece can offer a conceptual framework through which such questions can be investigated.
Various institutional changes may enhance the signals that disinterested voters perform. Empirical evidence that referendum voting increases voter information and engagement (Tolbert et al. Reference Tolbert, McNeal and Smith2003) suggests that implementing more direct-democracy style referenda has the potential to improve voter deliberation, including the deliberation of disinterested voters. Other potential reforms such as using epistemic conditions (Brennan Reference Brennan2016: 204–230), or rewarding voters based on how frequently they vote, could also improve voter deliberation and its effects. One radical alternative is to apply the juror model to voting. The idea is to randomly select citizens from the relevant jurisdiction and have the two sides present their case, for, or against, the proposition. The prolonged and deliberative nature of jury-style deliberations lends itself to an atmosphere of increased information and reflection. These are speculative proposals, but they illustrate the enormous range of institutional options for increasing the beneficial effects from disinterested voting.
Political exchange may be a legitimate function of voting, but it is not its only function. Mechanism design is a potentially powerful tool for achieving better political outcomes, but voter preference maximization should not be its only goal. The judgements of disinterested voters carry important information relevant to improving political outcomes. Thus, the question is, how can we structure political choice that optimizes the use of both types of information? Future voting design proposals should also consider the role that disinterested voters play and the broader social purposes that voting serves.
Acknowledgements
I’d like to thank Rick Jeffers, Scott Jeffers, Alexander Schaefer, Emilia Suggs, James Stacey Taylor, Christopher Freiman and Glen Weyl for their insights, comments and feedback for this piece.
Matthew Jeffers is an Independent Scholar working at the intersection of philosophy, politics, and economics. He obtained his Bachelor’s Degree in both Philosophy and Economics from University of Wisconsin-Madison and his Master’s Degree in Philosophy from Georgia State University. His work has appeared in philosophy journals and interdisciplinary economics journals. His primary research focus is on ethical issues that emerge at the institutional level.
