Is there an Alternative to Representative Democracy? The Promise of Lottocracy
No one pretends that democracy is perfect or all-wise. Indeed, it has been said that democracy is the worst form of government except all those other forms that have been tried from time to time.
—Winston Churchill
The appointment of magistrates by lot is thought to be democratic, and the election of them oligarchic.
—Aristotle
It is common sense among most Americans that electoral representative democracy is better than any alternative political arrangement. Moreover, most believe that the only legitimate alternative to electoral representation is some form of direct democracy. However, direct democracy - we are all told - would lead to bad policy. But Alexander Guerrero argues for a new political system in his paper Against Elections: The Lottocratic Alternative. This article is a summary of that paper.
The Perils of Electoral Representation
Guerrero uses two criteria in order to evaluate a political system: responsiveness and good governance. A responsive government listens and changes to what the people actually believe, prefer, or value. Good governance is government that leads to good outcomes. So how does representation stack up?
Accountability
In theory, elected representatives achieve responsiveness through accountability. Accountability is achieved through free and fair elections. Accountability also requires that ordinary citizens are capable of engaging in informed monitoring and evaluation of politicians. Unfortunately ordinary citizens are commonly ignorant about a politician’s conduct, about particular political issues, whether a politician is helping the common good, or even if a politician is helping serve a citizens’ personal interests. If citizens are ignorant of what a politician is doing, they cannot hold politicians accountable.
Therefore, elections can only achieve responsive and good outcomes only if citizens can meaningfully hold their politicians to account.
Representative Capture
It is still possible that representatives can lead to good outcomes even if they are not responsive to the ignorant public. Representatives have the time and resources to research issues, deliberate, and consult experts. They are also selected because they appear to be competent by the public.
Unfortunately, it is also possible that elected officials that are not accountable will be used to advance the interests of the powerful. The powerful can capture representatives by:
Influencing elections - determining who can be a viable candidate.
Influencing the elected - determining the political positions of politicians.
There is substantial empirical work suggesting that capture is a significant problem. In the book Affluence and Influence: Economic Inequality and Political Power in America, Martin Gilens demonstrates that US policy is mostly responsive only to the highest-income Americans. Jacob Hacker and Paul Pierson argue that the main source of the increase in income inequality over the last thirty years is due to the capture of US politics by the economic elite. Colin Crouch makes a similar case with UK politics. Lawrence Jacobs and Robert Shapiro argue there was a decrease in political responsiveness over the last decades of the 20th century, as a result of incumbency advantage and the proliferation of powerful interest groups.
What about the Good Ones?
It is possible that a well-intentioned class of people run for office, get elected, and do what is best for everyone even without accountability. Unfortunately, if these do-gooders act in ways contrary to the interests of the powerful, then the powerful will use their resources to either buy off these people or to crowd them out and replace them.
Proxies to Actual Knowledge
Even if we are utterly ignorant about the day-to-day productivity and management of our representatives, aren’t there some “work-arounds” so that we can still hold our representatives accountable? There are some strategies available:
Membership of political parties and attendance of their events.
Relying on endorsements from activists, media institutions, or public figures.
Unfortunately there are problems to relying on proxies. It is very difficult and time consuming to determine which proxies are credible, particularly if one wishes to find reliable proxies on many different issues. The effort of finding good proxies is almost as challenging as doing the research yourself. Moreover, some proxies will be too coarse-grained to help with accountability on specific issues, or others will be too fine-grained to save individuals time and effort. Finally, it is possible that no good proxies exist. Low profile issues might not attract interest groups to do investigations. High profile issues will attract powerful interests who shape and obscure the available information.
The more information intensive a political problem becomes, the less we are able to meaningfully hold politicians accountable. Therefore the more complex a political problem becomes, the less able elections will be able to bring about responsive outcomes to that problem.
Elections Might be Good for Some Things…
Guerrero also considers that electoral systems will do poorly for many political problems, but not all political problems. For example, Artya Sen empirically argues that electoral systems prevent famines. It is simple for constituents to understand that they are starving and famine is bad, and therefore they can provide appropriate electoral feedback.
Can We Fix Elections with Reforms?
There are many obvious proposals to improve the election system by imposing regulations on campaign finance, post-electoral employment, lobbying, election media coverage, etc.
Some also propose imposing strict term limits. This assumes that people are very good at selecting good politicians in elections while simultaneously believing that people are very bad at providing electoral feedback.
Unfortunately, none of the proposed solutions address the problems stemming from Complexity and Opacity. Many political problems are complex, technical, and information intensive. It is difficult to ascertain whether a politician has actually tried to act to the benefit of citizens, particularly not within the time between election cycles.
There are some ways to address Complexity and Opacity. For example, one can limit the size of the jurisdiction, such as “town hall” democracy, making it more possible that a smaller number people could become more informed.
Alternatively we can attempt to substantially improve our education system. Yet even if we provide free university education to everyone, there will be a limit on how individual people keep informed about the wide variety of political issues and what exactly their representatives are doing, even if individuals are well educated. Simply addressing Complexity does not address Opacity. One might understand the technical details but still not know enough about what their politicians are doing to address the problem. Moreover, there are too many issues and each one is too complex. Finally, there is little incentive to stay personally informed.
The Verdict on Electoral Representation
Elections will not tend to bring about responsive or good outcomes for issues that are complex. Yet we might still agree with Churchill - democracy is the best among the worst, despite the shortcomings of elections. In the next section, Guerrero proposes a new way to conduct democracy using a lottocratic system.
The Lottocratic Alternative
In lottocracy, also known as sortition, political officials are selected by lottery rather than election. There is some historical precedent for this method. For example in ancient Athens, lotteries were used to select government officials in many of its institutions. Lotteries were also employed to select leaders in Italian city states. Finally in modern times, Citizens’ Assemblies have been employed in Canada, the UK, Ireland, France, and around the world as advisory panels to their respective governments.
There are several ways to introduce democratic lotteries to government with varying degrees of ambition. Lotteries could be used to supplement an existing legislature. Lotteries could be used to completely replace a legislature. They could be used only for specific political problems (perhaps triggered by a special referendum). They could be used as an oversight mechanism, charged with making recommendations or having veto power.
Guerrero focuses on using lotteries to completely replace a legislature as a thought experiment. His test system is envisioned as:
A legislature is assembled by lottery.
Every legislature evaluates only a single issue (each focusing on for example, Agriculture or Health Care), rather than a single generalist legislature.
The members of the single-issue legislatures hear from a variety of experts on the relevant topic at the beginning of each legislative session.
Guerrero imagines that each of these single-issue lottery-selected legislatures (or SILL) consists of 300 people chosen by random lottery from the adult citizens of a country. Each person would serve a three year term. Terms would be staggered so that every year, 100 new people are chosen. All adult citizens would be eligible for selection. People would not be legally required to serve if selected. However, people would be given substantial payment to serve, and accommodations would be made for family and work schedules. Finally a civic culture may need to be developed to view service as a civic duty and honor. There would be a mechanism for removing people for bad behavior, such as failing to attend meetings, speaking out of turn, or showing up intoxicated. This mechanism would be structured to protect those who are unlikable or have divergent views.
Each legislature would meet for two sessions each calendar year. Each session would consist of phases of:
Agenda setting
A learning phase with expert presentations
Community consultation
Deliberation and discussion amongst one another
Drafting, revising legislation
Voting
Guerrero prefers these single-issue assemblies not to handle all legislation, but only problems that are either technical or are too dangerous for politicians to touch.
Members of these lottery legislatures would represent the greater public not from responsiveness but instead as descriptive and proportionately representative. Whereas elected representatives are expected to conform to the will of the electorate, the descriptively representative conform to the will of the electorate, because they statistically the same people that compose the electorate.
However, these lottery legislatures do not have in mind the idea that they must represent some particular constituency. Instead, each individual member of the lottocracy comes to have certain views about an issue after hearing from experts and engaging in consultation and deliberation.
Organizing the Stages of the Lottocratic Legislature
Agenda Setting
Lottocratic legislatures would decide on what to work on in the next session by a process of agenda setting, where a wide range of options are narrowed down. This gives time for the next legislative session to inform and involve experts and community members.
Expert Presentations
For each item on the agenda, the lottocratic legislature would hear from experts providing general background and specific information relevant to the agenda. As an example, in the “learning phase” of the British Columbia Citizens’ Assembly, experts made presentations about electoral systems. These experts also assigned a textbook on electoral systems as background reading. Political science graduate students were also hired to facilitate group discussions.
The motivation behind this system is to blend the virtues of expertise with the virtues of of policy-making by ordinary people.
What would the process to determine who the experts are? Some sort of qualification and selection process will be needed. There is also the possibility of expert capture by special interests.
Consultation, Deliberation, Drafting, and Voting
After hearing from the experts, the lottocratic legislature would begin the process of developing and deciding legislative proposals. Guerrero highlights some key elements of his envisioned system:
The process should include consultation with nonmembers, in order to gather information from the community as well as inform nonmembers about the issues and proposals.
The lottocratic legislature would then convene for deliberation and discussion. After discussion, the legislature would work together to draft proposals. This could be modeled on how legislative drafting works in other legislative bodies. Drafting aides and consultants could be on hand with expertise in drafting.
The Promise of Lottocracy
Unfortunately as the system as specified by Guerrero has never existed, it is difficult to assess. The virtues and vices are therefore speculative.
The Virtues
Corruption
Lotteries excel at preventing corruption or undue influence in the selection of representatives. Special interest capture of lottery members would be considerably more difficult. Lottery members do not need to run for office, so there is no way for powerful interests to influence who becomes a member. Without need to raise funds for reelections, it is also easier to monitor lottery members to ensure they are not receiving funds from powerful interests during or after service. Lottery members also rotate out of service regularly, so special interests would be forced to attempt to buy off new members again and again. It would not be possible to capture politicians who then become electorally unbeatable, and count on them as an ally for several decades.
Some may worry that powerful interests would just shift their focus from capturing elected officials to influencing the views of the populace as a whole. Fortunately, attempting to influence mass opinion is far more expensive than attempting to influence a politician or a single election. Moreover, citizens are limited in their “information retention” abilities. Many citizens just will not be paying attention to political issues prior to random selection.
Representativeness
Lotteries are more likely to result in a legislature that looks like every-day people from all walks of life. These legislatures are more descriptively representative than elections. Random people are more likely to be ideologically, demographically, and socioeconomically representative of the public compared to politicians capable of winning elections. For example, 44% of US Congressmen have a net worth over $1 million, 82% are male, 86% are white, and more than half are lawyers and bankers. Lotteries are also more likely to include people with greater range of life experiences and vocational skills, which could improve the quality of the outcomes due to improvements in cognitive diversity.
Lack of Political Ambition
Unlike elected politicians, the lottery selected generally have not sought out political office. The politically ambitious might have aspirations of power over others. Lotteries therefore may select those who would rather not have power over others. The politically ambitious might be easily corruptible, willing to exchange their principles for personal and electoral benefits.
On the other hand, some politically ambitious care about making things better and working on behalf of others. Lottery selections means you lose out on this kind of self-selection.
Long Term Thinking
Elections tend to encourage politicians to focus on short term problems which they can claim credit for address, while ignoring problems with longer horizons or solutions where it is more difficult to claim credit. Take for example climate change. As the worst effects of climate change will not be realized for decades, politicians tend to push back action again and again. Individuals chosen at random, however, are not hamstrung by these kinds of incentives.
Since the publication of Guerrero’s paper in 2014, three Citizens’ Assemblies on Climate formed by random citizens were conducted in Ireland, the UK and France. In all three of these assemblies, the citizens overwhelmingly and broadly approved of drastic action to protect the environment and take measures to combat climate change, including carbon taxes, meat and agricultural taxes, petrol taxes, bans on automobiles, and government incentives for green infrastructure. In all three cases, the actions recommended by citizens were either ignored or watered down by politicians.
Political Equality
Lottery selection better respects the ideal of political equality. Elections allow us to choose someone to rule over others. And though we have a right to vote, few of us have actual political power to change policy. In contrast to democratic lotteries, few of us have anything close to an equal chance of obtaining political power.
In comparison with democratic lotteries, some of us will have more political power than others in the short term. However this is far less pronounced compared to the election system due to rotating terms, and it may not even be true when measured over lifetimes, as our chances of having served at least once in our lives increases.
Democratic lotteries therefore come closer to realizing actually equal political power compared to any elected system.
The people chosen by democratic lotteries may also have a “humility of the chosen”. As Barbara Goodwin suggests, “those allotted high office would comport themselves more humbly… no one could boast of his/her elevation or advancement as being personally merited.”
Concerns about Lottocracy
Competence
The most common worry of democratic lotteries is competence. Wouldn’t entrusting decision making to a randomly selected body of citizens be a disaster, much worse than elections? Electoral politics ensures that the selected are at least somewhat intelligent, socially competent, and hardworking, right?
Guerrero summarizes a variety of responses that can address this concern.
Competence could be raised by creating incentives to get citizens to participate, or setting minimum thresholds (for example requiring a high school diploma), or improving public education. Thresholds would be controversial, as they would undermine some of the advantages of lotteries due to descriptive representativeness.
One could treat this question as a quality threshold and argue that most citizens would be competent.
One can also argue that elected officials might be more competent in one sense, but this is outweighed or undermined in the ways they are biased.
These arguments can be resolved empirically. It is not obvious that the average member of Congress is better able to understand technical policy issues.
In addition to Guerrero’s recommendations, it should be noted that a lottery body could be independently trained before service, for months or even years if so desired. If needed, each lottery member could be educated with an elite university education, at an expense far less than educating the entire public.
Experts
Another worry is that laypeople will not be sufficiently intelligent or educated enough to make good policy, even if they hear from experts - “perhaps especially if they hear from experts”. Alternatively, we may worry that citizens would defer to experts too much. Guerrero suggests we ought to monitor future lottocratic experiments to ensure the competence of normal citizens.
Since Guerrero’s 2014 paper, many experiments in lottocracy have been performed around the world. This includes America in One Room, the Climate Change assemblies of Ireland, the UK, and France, as well as Irish Citizens’ Assemblies considering abortion and gay marriage. The vast majority of evidence suggests that citizens were widely open to changing their opinion and open to listening to experts. For example in America in One Room, Republicans softened the views on immigration. Support for forcing undocumented immigrants to return to their home countries dropped from 79% to 40%. Support for more visas for low-skilled workers increased from 31% to 66%. For Democrats, support for raising the federal minimum wage to $15 dropped from 83% to 59%.
Guerrero’s worry that citizens would defer too much to experts has not be thoroughly investigated.
Guerrero does not go into detail on how experts would be selected in his lottery system. Arash Abizadeh for example has suggested that the experts should be selected by a combined lottery-selected and election-selected bicameral legislature. Alternatively the experts can be selected solely by a lottocratic legislature similarly to how legislatures select bureaucrats.
Participation
Though citizens would have an equal chance to participate in lottocratic legislatures, that probability might be small in a large country. Therefore democratic lotteries might possibly limit political active roles ordinary citizens could play.
However, political participation right now is already very limited, with most people not bothering to vote. Participation rates in local elections may be around 10%. In democratic lotteries, citizens would still be allowed to play an active role during the Agenda Setting and Community Consultation phases.
Conclusions
Guerrero believes that despite potential deficiencies, democratic lotteries could perform better than elections in two important dimensions: responsiveness and good governance. Guerrero recommends a piecemeal, small-step, small-scale approach to introducing democratic lotteries into our institutions.