Should Democracy Work through Elections or Sortition? *

Political Equality

In considering political equality, I look first at descriptive representation and then at the influence of unevenly distributed wealth on the political system. Both bear on the argument for sortition, but in very different ways. ⁸

Popular Control

In many ways, the heart of democracy is the ideal of popular control and accountability. When a political system’s scale is so large that the people cannot directly govern themselves, democracy requires representatives to act on the public’s behalf. For this to work properly, accountability is essential. So let us now consider the strengths and weaknesses of elections and sortition in this regard.

For an electoral system, the major limit to accountability stems from the independence of elected politicians from their constituents when making decisions. One problem is that elected politicians have little to no incentive to respond to constituents from other parties. Moreover, a vote for a candidate—or even for a party—is a blunt instrument to signal one’s complex policy preferences. Party discipline, which candidates owe to the party that then steers their votes once in office, further limits a candidate’s responsiveness to constituents. Finally, most policymaking involves both opacity and complexity, ²¹ which makes it difficult for constituents to grasp the consequences of representatives’ votes on their behalf.

Nonetheless, a well-functioning electoral system still provides a certain degree of blunt accountability. ²² Although party discipline does restrain individual representatives, it also allows for accountability in a collective sense by transforming platform promises into policy. Political parties can also tackle complexity and opacity by foregrounding their broadest contrasts with opposing parties. At their best, parties give constituents real choices in terms of broad values and policy priorities. Elections then give voters the chance to hold parties accountable for delivering, or failing to deliver, on those promises. Indeed, anyone doubting the significance of such choices is insensitive to the consequences of electing, say, Donald Trump as opposed to Hillary Clinton.

What about sortition? Since members are not elected, there are no direct mechanisms of accountability in the usual sense. However, Philip Pettit has argued that sortition does provide a kind of popular control, which he refers to as “indicative.” ²³ To borrow Pettit’s example, if I want to have some accountability over, say, a new university committee that has been convened to investigate how to make philosophy more appealing to female students, one way to do this is to have the committee members run for election. But another way is to establish a system (such as sortition or stratification) which ensures that some of the members are “like me” in that they share my values and principles and are likely to make the same decisions I would make if I were on the committee. If such people really are similar enough to me, then Pettit is right that ensuring their presence really does give me some influence and control over what happens. The argument is similar for a sortition body: if the descriptive representation works well so that some of the members are “like me” in their values, then there is indeed a kind of popular control here. Granted, I do not have any direct control over the decision makers, but if I can ensure that the process includes people “like me,” then I do gain some indirect control over the decisions made.

The difference between elections and sortition is that although both provide popular influence over representatives at the initial time of selection, as time goes by there is no way for regular people to continue to exert real influence over sortition members. Citizens can try to participate at a public consultation, but only a small number of members will be able to listen to a very small proportion of the citizenry, and even those citizens who do get the chance to participate have no power to make members listen; sortition members are always free to ignore them without consequence. The difference with elections, of course, is that they give representatives continual incentives to be sensitive and responsive to the desires of their constituents. Another problem with sortition in this regard is that in such a system regular citizens do not get to participate in the formal political system at all. A vote once every few years is not a lot of political participation; but it is meaningful. Under sortition, there is a sense in which the people would be disenfranchised from the political process. What is particularly worrying here is the lack of clear, formal channels for citizens to transform their dissent into political power and the frustration this might generate. If citizens cannot collectively mobilize through elections to get what they want in a legal, nonviolent way, the incentive to look to extraconstitutional means becomes that much greater.

Another issue relevant to the idea of accountability is whether the deliberations and the final vote should be public or held in secret. There is a difficult trade-off here: secrecy may improve the quality of deliberation, as it makes it easier for members to give up old positions and change their minds; secrecy of final votes may also help with the problem of corruption, since prospective bribers would be unable to verify how any representative voted; secrecy can also protect the decision makers from the embarrassment of being shamed or ridiculed (a real prospect in the age of social media). On the other hand, the more decisions are made in secret, the less accountable the decision-making process.

There is no easy answer here, and I would not pretend to have total confidence in the solution, but my considered judgment, contra Gastil and Wright, is that while occasional deliberations may be confidential, final legislative votes should probably be public. The central weakness of sortition is its lack of accountability, and we should not exacerbate the problem. It seems hard to believe that the general public could accept having major decisions made in secret. That would mean having a legislature where not only can we not throw the scoundrels out, we cannot even know which scoundrels are making the decisions that are affecting us. That strikes me as a step too far. Allowing sortition members to stay barricaded behind the walls of secrecy does protect them from ridicule, but it also removes them too far from the push and pull of the public sphere. In a democratic society, accountability requires transparency: the right to know precisely who is making decisions, to look them in the eyes and demand the reasons for their actions. An important aspect of accountability is lost if we cannot expose actual human decision makers to public scrutiny and contestation and put pressure on them to take account of what the people think.

If the decision makers are known, then members of the public can try to exert the pressures of public discourse. The public can try to persuade, convince, shame, encourage, support, and morally exhort sortition members (although threats, malicious slander, and hate speech should be illegal). Decision makers will know and feel that their final votes are being watched. ²⁴ If, on the contrary, the decision makers are unknown, then activists cannot communicate with them, cannot write letters, cannot invite sortition members to participate in public debates, cannot send representatives of social movements to reason with them, cannot hold protests or vigils outside their office, and so on. In this way an important avenue of accountability is lost.

Thus sortition members should have to account to the public and media for their votes. That would certainly be an intimidating thing for regular people to do. But note that the pressures on them are somewhat different than on elected politicians. Before a vote takes place, it would be completely acceptable for a sortition member to avoid media questions by simply saying, “I do not yet have a firm opinion; we are still learning and deliberating.” That is legitimate in a way that it would not be for elected politicians, who are always supposed to have a confident answer to every question. This fact would significantly ease potential embarrassment. After the vote, however, sortition members should have to face the music and explain to the public why they voted the way they did.

In sum, elections offer blunt accountability—perhaps more via parties than individual candidates. Sortition gives the public a kind of control over the selection process, but it lacks the disciplining function provided by elections. Making sortition members’ final votes public would somewhat help with accountability, although it also risks opening up members to ridicule and censure. All in all, elections do better on this score.

Deliberation and Impartiality

Ideally, a democratic body should make decisions through high-quality deliberation. The people’s representatives should be honest, thoughtful, open-minded, impartial, and public-spirited—asking not “What is best for my group?” but “What is best for all of us?” A chamber populated by such persons would establish consensus where possible, fair compromises when appropriate, and always let the unforced force of the better argument prevail.

How well do elections serve this end? One problem is that the public that does the electing is often extremely uninformed. ²⁵ Moreover, inherent logic of electoral competition undermines the possibilities for good deliberation. There are at least four reasons for this. First, the skills and traits most useful in elections are in many ways the opposite of those of good deliberators. Second, electoral competition creates a strong and continual incentive to “score points”—never to give one’s opponents “a win,” even if doing so would serve your own constituents’ interests better. Indeed, few things can deflate one’s enthusiasm for democracy more than watching parliamentary discussion, with its incessant mudslinging, booing, clapping, and stomping. Third, electoral competition leads representatives to focus on short-term solutions, for which they can claim credit. ²⁶ Such competition also stunts political learning, since veteran politicians can be punished by party and voters alike for changing their minds.

What about sortition? We cannot simply assume that bringing people together results in good deliberation. Designing the conditions for high-quality deliberation is a difficult task—part art, part science, as Lyn Carson shows. ²⁷ Using this knowledge, it is plausible to envision a well-functioning (although imperfect) sortition chamber, divided into the major branches of public policy, where members engage in periods of learning from diverse experts, regular public consultations, and ongoing small-group deliberation. Skillful moderation and facilitation can foster relatively equal member participation and a respectful, caring atmosphere, especially if some discussions happen in closed sessions, free from the pressures to perform that come from publicity.

None of that will happen automatically; it will require a carefully managed infrastructure of resources and support, skillful facilitators, and administrative oversight. Since the administrators and facilitators must play a key role in the sortition body, we must carefully structure this background infrastructure (which we might call the Office of Deliberative Administration). It will need to be staunchly neutral on all ideological and policy questions and concern itself only with the practical matters of deliberation, such as procuring experts.

How exactly this office should operate remains an open question of utmost importance for future research, since it must not be allowed to influence the sortition chamber (e.g., by selecting experts from only one side of an issue) or be captured by partisan interests. ²⁸ One possibility might be to require the office to prepare regular reports of its activities, then appear before a committee of the sortition chamber to justify and explain its actions. The sortition members could have some authority to hire, fire, or reconstitute aspects of the office, since the sortition body itself has a vested interest in maintaining its legitimacy by showing itself to be a rigorously deliberative and nonpartisan space.

Advocates of sortition insist that a legislature by lot would perform significantly better than an elected chamber in terms of deliberation and impartiality. Without party discipline or the need to pander to any constituency, members would be free to listen to each other, learn, change their minds, and be guided by the force of the better (and, I would hope, more caring) argument. Moreover, the descriptive representation of the body means that it would be much more socially varied—encompassing the experiences of not just rich white men, but women, the poor, renters instead of owners, and employees instead of employers. As a result, it would be epistemically richer. Indeed, a wealth of recent experiences with minipublics shows that, under the right conditions, citizens can engage in high-quality deliberation. ²⁹

Beyond this, there is one serious caveat to the deliberative potential of a legislature by lot, namely, whether the long tenure of members would lead to the emergence of factions or parties, thereby undermining the quality of deliberation. Recalling the scale-transformation problem, minipublics have little to say on this issue, given their short lifespans and limited authority.

In bicameral sortition models like the one proposed by Gastil and Wright, there are several sources of pressure pushing toward factionalization. Elected politicians will have a strong interest in actively lobbying sortition members to join their party—if not formally, then at least as informal political allies. A long tenure means that sortition members themselves will have an incentive to self-organize into factions and coalitions, to be more effective in getting their preferred legislation passed.

On the other hand, one of the strongest motivations to form parties in the first place—winning elections—would be absent. The sortition chamber would have a deliberative structure of small group discussions, learning, and facilitated exchange of ideas. These encourage people not to stick to just one position, but to evolve in an open-minded way. In addition, the members themselves have at least some motivation not to form political parties or obvious coalitions, as their own legitimacy (and hence power) depends on being able to convince the public that they are impartial deliberators. Overall, it seems unlikely that a sortition chamber would become as rigidly factional as electoral chambers, so we can expect sortition’s deliberation to be somewhat better. Yet we should take note of the trade-off here: longer tenure leads to more competency (as members become expert in their various policy areas), but it may also increase the likelihood of rigid factions forming and deliberation deteriorating.

In sum, electoral systems rarely create the conditions for good deliberation. A sortition chamber has greater potential to deliberate impartially, with less posturing and factionalism. It is fair to say that advocates of sortition are justified in having some optimism on this score, although the lack of evidence of really existing sortition bodies means that it should be optimism of a cautious sort.

Competency

Every now and then, a politician will get elected who is strikingly incompetent. Donald Trump is a paradigmatic example, ³⁰ but such cases are the exception, not the rule.

Getting elected usually requires a long period—often years—of participating in local politics, working one’s way up the party ladder, hosting events and fundraisers, engaging in debates and interviews, persuading other party members to select you as a viable candidate, then convincing tens of thousands of strangers to trust you. This grueling process usually weeds out incompetent people. Political parties also play an important role in generating policy expertise. It is not necessary (nor would it be possible) for individual politicians to be experts on all of the different policy areas; belonging to a party provides politicians with a massive infrastructure of knowledge and shared policy goals. Thus, although we should not exaggerate the competency of elected officials, electoral systems tend to generate competency.

By contrast, this criterion poses a challenge for sortition. Will random members of the public prove capable of understanding and making sound decisions on complex policy problems? Consider two types of people who could be problematic for a sortition chamber, whom I refer to as the Unknowledgeable (e.g., a high-school dropout with a learning disability) and the Ideologue (e.g., a committed white supremacist or doctrinaire Leninist).

The Ideologue poses a manageable problem, as it would become obvious to other members that such an individual is closed-minded, unwilling to work in a deliberative spirit, or disrespectful. In extreme cases in which the Ideologue becomes altogether disruptive, members should be able to recommend their expulsion. In more common instances, the sortition body can work around such a person, or simply ignore them.

The issue of the Unknowledgeable sortition member is more difficult. It seems likely that most would not volunteer to participate in the first place. Those who do might be educated in general knowledge and helped to develop their capacity for judgment—but only partially. The sortition chamber faces a trade-off in that it could impose some basic competency requirements (e.g., basic literacy, or a high-school diploma) to prevent the worst problems of incompetency, but that would also reduce the descriptive representativeness of its membership.

In addition to the problem of the Unknowledgeable, another deep competency problem flows from the scale-transformation issue. Since a sortition chamber would be so much more complex than a minipublic such as a citizens’ assembly, even its more knowledgeable members may not prove competent—at least by comparison with the average elected official.

Consider some of the details about what an all-purpose sortition chamber would have to do. Every year there would be hundreds of bills to discuss, from very different policy fields, each with its own history and problems, and each requiring its own expertise. Moreover, bills from one policy area would invariably affect very different areas, which means that amending and voting on such bills will require competency not only in one’s own policy area but in all the connected areas. The problem is likely to be frequent, because one issue often interacts with others, and commonly—indeed very commonly—bills will affect each other because of budgetary constraints.

Almost every political issue affects the budget because policy solutions compete for the use of limited revenues. This is precisely where a citizens’ assembly diverges from a sortition chamber. Citizens’ assemblies have, for good reason, generally focused on the rare political issues that do not involve money, such as electoral reform or gay rights. This makes them vastly easier to handle competently because they do not involve weighing fiscal trade-offs with competing issues. But imagine trying to do a citizens’ assembly on any normal political issue, such as day-care policy, education, or environmental protection. How could lay citizens possibly decide what kind of day-care system to implement if they do not grasp the larger budgetary issues it involves? How could they come to a rational decision about whether it is better to provide expensive publicly provided day-care centers, or cheap tax credits partially to support families providing their own child care, without knowing the relevant trade-offs? Is it only possible to provide universal day-care by slashing welfare rates? Can we raise taxes on the rich, or what if we reduce military spending?

For this reason, I believe it is impracticable for sortition members with knowledge solely of their own areas to come to rational policy decisions. This is why proposals for single-policy sortition bodies are unlikely to work well. ³¹ Such bodies cannot deliberate meaningfully if they are barred from weighing the ramifications of policy solutions for other issues beyond their agenda.

This problem presents defenders of sortition with a serious design question. How could one envision a sortition body that enables competency? Others will have to take up this challenge, but I want to suggest one possible solution. Imagine that sortition members had a tenure of, say, four years, of which the first two years were training—without any legislative powers. The first year could involve training in budgets, debt, taxation, and distributive justice; exposure to the major fields of public policy and the functioning of government; and learning how to deliberate, with empathy, and with a sense of responsibility to the public good. In the second year, members could be selected by lot into one of the major ten or so fields of public policy (environment, health, military, economy, etc.), in which they would spend the rest of their training period “interning,” so as to develop familiarity and competency in the area about which they will spend the next two years making decisions. ³²

Such a body should be sized so that it can divide itself into departments big enough for diverse deliberation and remain statistically representative of the population. (As noted earlier, a larger size also limits bribery and corruption.) One possibility would be a 1,000-person body, divided into ten departments of 100 persons. Each department would focus on policy in its own area before submitting legislative proposals to the entire body. For proposals to become law, the departments would present the results of their deliberations, as well as their recommendation, before a general vote. It would be the job of the entire sortition legislature to weigh the costs and benefits of each proposal against those of other departments before making a final decision.

In sum, electoral systems tend to produce political competency as a result of the weeding-out function of elections and the intellectual support of parties. A full-bore sortition chamber faces difficult issues of unknowledgeable people and complexity arising from the cross-cutting nature of issues—almost all of which have budgetary implications. In theory, a sortition body might overcome such difficulties by having significant periods of prior training, being large enough to allow specialization, and retaining final authority in a larger body that can weigh the costs and benefits of various proposals. Nevertheless, humility in the face of uncertainty compels me to score electoral systems as faring better in the matter of competency.

Conclusion

Reflecting back on all four criteria, I see the main strengths of the electoral mechanism as accountability and competency, whereas its main weaknesses are generating political inequality (via descriptive unrepresentativeness and a systemic bias toward wealthy interests) ³³ and a systemic propulsion toward partisanship that undermines deliberation and impartiality. The virtues and vices of a not-yet-existing sortition chamber are more speculative. Nevertheless, such a body would surely perform better in terms of political equality because of its enhanced descriptive representation and its better insulation against the influence of wealthy interests. In addition, a sortition chamber would likely outperform its electoral counterpart at deliberation and impartiality, although getting lower marks for accountability and competency.

Three important conclusions follow. First, elections are not the only game in town. In many underappreciated ways, sortition has much to offer democratic theorists and practitioners.

Second, because neither election nor sortition by itself can satisfy the full range of democratic values, an optimal democratic system would need to combine both mechanisms, such as through the bicameral system advocated by Gastil and Wright. An additional reason for doing so is that elections and sortition each offer a crucial type of representation. In an elected chamber, the aim is to have representatives of the entire population take into account their actually existing interests. In such a chamber, discussion would ideally take the form of bargaining among fixed interests, among MPs who are highly monitored and revocable (playing a role as delegates, with limited independence). In a sortition chamber, by contrast, the aim would be to have a descriptively accurate sample of the population engaged in quality deliberation to learn what a representative sample would want in ideal deliberative circumstances (with members who are not delegates and so have substantial independence to change their minds). In other words, combining both mechanisms would allow us to profit from having representatives of our actually existing interests as well as our hypothetical postdeliberative interests—both of which are valuable, and neither of which we should want to do without.

Third, it is likely that a sortition chamber will lead to more progressive policy than an elected chamber, because there would be less elite representation in government (e.g., fewer bankers and lawyers, more caregivers and workers) as well as less systemic bias toward money.

Our comparative analysis has also sought to identify the trade-offs that would exist in building a sortition body, which all future designers will have to grapple with. These are as follows:

In addition to these trade-offs, there are also a number of open questions for future research into sortition:

In conclusion, contemporary democrats are wrong to simply assume that democracy requires elections and only elections. Our democratic values cannot be well satisfied by the electoral mechanism alone. Nevertheless, we should not leap to the false conclusion that elections are worthless. Neither electoral fundamentalism nor abolitionism is an appropriate response to our complex political situation and clashing democratic values. Sortition has significant democratic potential, and democrats should be open to exploring it. However, the largely untested nature of a national sortition body (and subsequent uncertainty regarding its virtues and vices) leads me to conclude that a piecemeal, small-scale, step-by-step approach to introducing sortition bodies would be wise.

Overall, democrats should be more willing to consider implementing sortition mechanisms than they commonly are. Being a democrat means having a number of values, but one of them is the belief that regular people, in good situations of nonviolence, learning, support, and deliberation, can arrive at quality political decisions. In this sense the radical democrat Ella Baker was right: give the people a light and they will find a way. ³⁵