As many of the essays in this symposium suggest, electoral rules have a dramatic effect on political outcomes. These rules also enable us to achieve different normative ends such as maximizing the likelihood that every citizen votes for a candidate who wins, maximizing the correspondence of policy outcomes to citizen preferences, and ensuring an ethnically diverse legislature, among many other desiderata. Since electoral rules do not equally achieve each of these goals, and since some of these aims are incompatible with some others, the choice of an electoral system will have normative consequences. For instance, in choosing an electoral rule, we may end up having to sacrifice the goal that every citizen votes for a candidate who wins so as to maximize the likelihood that policy outcomes correspond to majority preferences.
Here we seek to highlight the normative implications of electoral reforms, while recognizing the tradeoffs that instantiating our commitments in political institutions necessarily entail. To preserve even a core value such as political equality in an electoral system may entail unacceptable sacrifices of a competing value—or even a sacrifice of equality on a different dimension. As such, engaging in political experimentation and innovation may be our best hope of expanding the range of institutions available to realize these commitments and of clarifying where our normative priorities rest.
Democracy is committed, in the first instance, to political equality. In ancient Athens, this entailed the equal right to put oneself forward to hold political office and the equal probability, via the lot, of being selected (for most magistracies). Yet it is far from clear what equality among citizens ought to mean in modern representative democracies. Under universal suffrage—one person, one vote—each citizen in principle ought to have equal opportunity to choose the candidate who she believes will best protect her interests. Robert Dahl has famously defined democracy by reference to the “responsiveness of the government to the preferences of its citizens, considered as political equals.”Footnote 133 But note that already there are potential tensions between the equal capacity to choose a candidate, and the ability to have one's preferences count equally: my preferred candidate may have no chance of winning; my candidate, even if elected, may be in the position of a permanent minority and have no means of realizing my policy preferences; and even if elected and in the majority, my representative's votes may not always track my policy preferences.
Even if we specifically aim to improve political equality in one form, we may end up increasing inequality measured in a different fashion. As Charles Beitz has suggested, we might decide that political equality should obtain primarily in reference to the legislation created rather than to voting or to the likelihood of electing a candidate.Footnote 134 In that case, a commitment to proportionality may or may not lead to a greater chance of achieving legislative outcomes. Greater homogeneity within the district may increase the capacity of constituents to choose a representative who shares their values and interests, but at the cost of having a representative who may be marginalized once she enters the legislature. This is what David Lublin nicely titled the “Paradox of Representation.”Footnote 135 The alternative that he suggests is to have minorities remain within single-member districts at a large enough level to affect some moderating influence on the resulting election, even if their numbers are not large enough to elect their own. Although Lublin was discussing only minorities, the point can be generalized to party affiliation as well, were small parties to be represented proportionately in legislatures. Maximizing political equality in the sense of having representatives who reflect the distinct interests of particular populations may lead to a loss of equality measured in terms of political outcomes.
Although achieving a perfect harmony of values is unrealizable, there are dimensions on which our existing electoral institutions fall short and might be subject to improvement. It is worth exploring whether serious institutional innovations might help us to realize our normative aims. History provides evidence that imaginative ideals have led to lasting political changes. When Thomas Hare proposed a proportional system as an alternative to majority rule to elect representatives for the English House of Commons, it was derided for being absurdly complicated because it required every voter to rank order hundreds of candidates on a list. Who would have predicted that his ideas would lead to the single-transferable voting system (“preference voting”) and other forms of proportional representation now in practice today, thanks in part to John Stuart Mill's own account of Hare's system in Considerations on Representative Government? Today organizations around the world—from local non-profit groups to national governments—use a variety of these schemes to translate underlying voter preferences into the institutions of government.
Political theorists have recently focused attention on the incentives of representatives to act in ways that may fail to realize the interests of their constituents, or may promote the interests of their constituents at the expense of the national interest or “common good” more generally. Insofar as these issues entail the question of how best to ensure the equal consideration of citizens' interests, they are importantly connected to the issues of political equality addressed earlier. But they also transcend these issues to raise normative questions of the ethics of political representatives and the responsibility of citizens for monitoring their agents. Two institutional innovations designed to address these issues help us both to clarify both some of the deficiencies in current electoral systems and to reimagine the possibilities of electoral democracy.
Alternative Constituency Design
Nearly every democracy groups voters into territorial subgroups before counting their vote. This is an historical artifact and may be less important to large states than it was three centuries ago and as it continues to be in smaller communities. The original justification for territorial districts may have been the close ties voters within a district had with one another, a justification that was plausible when the average size of a colonial district was roughly 3,000 people. But today, in the United States for example, electoral constituencies for the House of Representatives number some 600,000 people, much larger than the original number of 30,000 set in the constitution in 1789. And that original number was on average 10 times the size of districts for colonial legislatures.
The fact that electoral constituencies are based on geographical lines creates incentives for representatives to support local spending as part of serving their constituencies. Sometimes this spending is also in the national interest. But at other times, most famously in the “bridge to nowhere,” locally-oriented spending—simply a way to secure votes—reduces resources available to projects of a more general interest. Political biases toward local interests are not natural. Rather, the institutionalization of representation along geographical lines—territorial districting—renders politics local. The alternative would be proportional representation in which electoral constituencies are essentially formed by voters who cast a vote for national parties. Other alternatives include functional representation and representing individuals by profession or other organized interests. These systems would not eliminate special-interest spending in the name of the national good, so much as they move local spending to spending on other interests on which the constituency is defined (e.g., interest group, political party, and profession).
Thomas Pogge has gone so far as to recommend “self-constituting constituencies” that would allow individuals to have their votes counted in whatever way they choose.Footnote 136 Such an alternative has the benefit of leaving to the group the manner by which they are represented, and to define for them what it means, for example, to be represented as a woman, even as it raises the likelihood that one of its own members is elected. Pogge's recommendation maximizes the value of autonomy by placing the choice of a constituency in the hands of voters. But autonomy at the level of constituency definition is likely to leave districts more homogenous since they will attract like-minded voters (each of whom wants to be represented by their party identification, or by their interest in whatever feature it is by which they would choose to be defined). This may generate the paradox of representation described earlier, making it more desirable to sacrifice autonomy in order to achieve better policy outcomes.
In contrast to territorial representation or more concentrated-interest representation, we can imagine electoral constituencies that are defined by the national interest, rather than the interest of a part. This could involve randomizing the assignment of citizens to permanent, involuntary, heterogeneous electoral constituencies, producing non-territorial districts in which each would look like the nation as a whole.Footnote 137 In a legislature of, say, 435 seats (the number in Congress) each voter would randomly draw a number between 1 and 435 (inclusive), and she would be a member of that constituency for the rest of her life. The random district would reduce partisanship, by having each representative pursue their own constituents' interests even as that would mean pursuing the nation's interest as a whole. In addition, the permanence of such districts would enable the same group of voters to authorize a representative to act for them and then reward or punish this person at election time.
The use of randomization is in part designed to promote the inclusion of diverse social groups. The membership of each constituency would be national, and demographically identical to each other, as well as to the nation as a whole. The median voter would not only prevail within a constituency but nationally. The hope, often implicitly, is that through capturing the diversity of the citizen body as a whole, the institutions will better reflect either the common good or at least not be subject to the systematic distortions and partiality associated with electoral systems (again, perhaps most importantly, wealth). The resulting policy implication is that issues for which there are clear national majorities would have constant support within Congress. But issues on which the nation is divided would be the primary ones on the basis of which representatives might distinguish themselves from each other.
Since the constituents of each district would be nationally distributed, campaigns would not find it effective to advertise using broadcast media, but only through direct e-communications, dramatically lowering the costs of campaigns and thus reducing the influence of media in politics. And in Congress, spending and other decisions would still be based on serving the good of a representative's constituents, but now the constituency's interests and the nation's interests would be identical.
Accountability and the Euthynai
That elections are insufficient for accountability is a familiar and justifiable lament. Monitoring one's representatives is very costly for voters, even aided by opposing parties, and representatives have an incentive to conceal bargains that constituents would reject. Further, the bluntness of the vote makes it imperfect as a mechanism of retrospective accountability for several reasons, among them that it entails both a retrospective assessment of the representative's performance in office, which is itself comprised of a multitude of votes, and a prospective judgment as to whether an opponent is likely to perform better.Footnote 138
In recent years, theorists have begun to defend the value of extra-electoral mechanisms of political accountability. Jeffrey Green has recently argued that the primary role of the people in contemporary democracy is to be “spectators” of their leaders, and, as such, the critical ideal ought to be candor: “the institutional requirement that leaders not be in control of the conditions of their publicity.”Footnote 139 He defends the increased use of public inquiries, in which leaders are subject to explicitly political investigation and trial, holding that they would encourage direct scrutiny while removing the power from the leaders under examination to shape the spectacle. John McCormick praises Machiavelli's defense of political trials, and urges contemporary republicans to endorse them as a means of improving the accountability of representatives and of fostering citizens' control over the wealthy elites who today dominate political offices.Footnote 140
Though Green in particular offers suggestions for how some of the obvious risks associated with political trials might be mitigated, nonetheless, we might still not wish to develop a political institution that could so readily be subject to partisan abuse. Nor would we necessarily want an accusatory framework for such an accountability mechanism. Yet the public nature of rendering accounts, especially given the incentives on the part of incumbents to obfuscate features of their record, might well have benefits. Ancient Athens offers a model of one such mechanism. Magistrates and other public officials in Athens were subject to scrutiny after their term in office, a mechanism of “rendering accounts” called euthynai. A first stage was focused on financial improprieties, but the second phase entailed a mechanism by which any citizen or metic could present a written accusation of malfeasance; if the “corrector” for the magistrate's tribe deemed the charge justified, it would be handled by public or private prosecution.Footnote 141
One might suggest that such an institution is unnecessary, because opposing candidates will have an incentive to monitor and disclose evidence of wrongdoing or simply votes that are at odds with the public interest, as construed by the competing candidate. However, because the opposing candidate herself has a strong incentive to characterize the record in a negative light or to misrepresent the incumbent's position, there is reason to think that the voters would not, and should not, necessarily trust the challenger's presentation of the incumbent's performance. As such, and given the limited incentives on the part of the incumbent herself to provide a full rendering of her performance, an additional mechanism might be necessary.
Imagine a public hearing at which the incumbent would present her voting record as well as perhaps other evidence of her performance in office, and would be subject to challenge on this record by a panel of judges. Though it is surely the case that this would expose the incumbent to substantial criticism, which might well benefit the challenger, it is not obvious that this is unattractive; the incumbency advantage is so strong that mitigating it on the basis of a careful evaluation should not raise serious concerns. Further, to the extent that the prospect of this scrutiny encourages closer relationships with constituents and public reason-giving for unpopular decisions, the performance of representatives might be enhanced.
One might ask how different the euthynai is from what the free press already does in contemporary society. Journalists regularly compel politicians to render accounts in public forums such as press conferences and televised interviews. Citizens also participate in this process by submitting questions, a process that the growth of the internet has facilitated. Yet the profit interests of the media and its penchant for spectacle frequently reduce the process of rendering accounts to the exchange of sound bites. The public emerges more entertained than informed. Nonetheless, as Green would argue, stripping away the leaders' own power to orchestrate the euthynai makes it a spectacle worthy of public observation.
Conclusion
Choices between electoral rules entail compromises and tradeoffs among different values—and even the particular manifestation of these values, as the discussion of political equality suggested. Within democracies, there is likely to be disagreement about the relative weight that should be ascribed to these values. Because of the technical nature of electoral systems, citizen engagement in electoral reform tends to be relatively limited—except, perhaps, where redistricting is concerned. But, as we hope to have shown, the normative consequences of electoral design are significant.
Though there are serious costs associated with instability in an electoral system, there are also substantial liabilities to sclerosis: electoral reform in many ways constitutes an ongoing process. Remarkably, perhaps because of the desire to enable electoral rules to remain flexible enough to cope with important demographic or political changes, electoral rules are rarely given constitutional status. Although the flexibility of electoral rules may generate an incentive for temporarily dominant parties to lock in advantages, it also provides citizens and their representatives an opportunity to ensure that electoral rules reflect their normative commitments—indeed, an opportunity to reflect on the nature of these commitments and their limits. The realization of our ambitions for democracy more generally, however, may require us to think creatively about institutional innovations beyond elections.
