When most people think of representative democracy, they usually think of a system in which people vote on who they want their representatives to be and then those representatives create and vote on policies. However, an alternative system that makes use of representatives and is, at least under some definitions, democratic is sortition (also known as demarchy or lottocracy). Sortition is a system in which the representatives that craft and vote on policy are selected at random from the population. In this post I’ll be focusing on comparing the use of these two systems for the legislative branch (not executive or judicial positions). I have a fair amount of skepticism about the prospects of sortition but I’ve recently become more interested as I’ve thought more about the flaws of election-based representative systems.
If you’ve seen some of the other posts on this blog, you could probably guess that I think more consideration should be given to how elections are conducted in representative democracies because it does seem like many of design details have the potential to significantly impact the performance of our institutions. To take one example, changing what voting methods are used seems like it could change how much people’s votes reflect their true preferences, how powerful political parties are, and how proportional the representatives selected end up being of the overall population along various dimensions. However, there does seem to be at least one major problem that any election-based representative system will run into, which is that as the population gets large the probability of any individual vote being pivotal rapidly approaches zero. It seems like it will always be challenging for this not to lead to problems of rational ignorance. That is, if the probability that your vote will affect the outcome is negligible, then someone could reasonable conduct a cost-benefit analysis and decide it’s not worth the effort to become well informed about all of the candidates’ positions and the effects of policies. Of course, some people enjoy thinking about and engaging with politics but even then there may be the potential for “rational irrationality” in the sense that if the main consequences of holding certain political views are primarily social or how they get to think about themselves then they may be biased to approach issues differently than if the main effects were actually changing policy.
This seems like one area where sortition does have an advantage, since while there would still be some voting among the randomly selected subset, perhaps it could be small enough to have the expected value of each member’s decision be sizable in general. Any kind of pure sortition based political institution would be so different from are current system that it’s hard to be able to anticipate all of the potential problems or benefits but here are a few others that seem like they are worth considering:
As I indicated before, election-based representative democracies can vary significantly in terms of how well the elected officials end up representing the population ideologically or along other dimensions. It seems like sortition could offer stronger representativeness guarantees, since with a reasonably sized legislative group (say several hundred people) a basic law of large numbers argument could guarantee that with a high probability the proportion of people who, for example, support higher taxes or are under 55 would be approximately the same as in the population they were sampled from. It would be reasonable to have some rules restricting who could be selected (e.g. not young children) and this choice would have important consequences for the system similar to the choice of who can vote in the election-based version.
If we assume that the randomly selected people are given the option to decline then we would likely want to make sure that people are not financially or legally penalized for participating. Part of this would plausibly be that they would paid well during whatever time period they’re working as part of the legislator. However, there would still likely be some selection bias in terms of who accepts and by paying them well this would then change them to make them unrepresentative of the population in a pretty important way (at least while in office). Of course, this same unrepresentativeness also exists, usually being even worse, with the election-based system but one could argue that since the poor can still vote those representatives will have more of an incentive to care about their problems.
This leads to one of the main concerns people have about sortition, which is that since the representatives don’t have to worry about re-election they are less accountable to the general public. This positive incentive for elected politicians certainly does exist to some extent but it seems like the positive effects here can be considerably undermined by the rational ignorance/rational irrationality effect mentioned earlier. Additionally, it may also lead to perverse incentives (e.g. when politicians can be more successful by prioritizing short-term political gain over tackling long-term problems). Also, if the reason why we care about this kind of accountability is the we want the politician to reflect the values of the voters then we should only be critical of sortition on this front in so far as it fails to achieve this end through other means (with the unrepresentativeness of incomes being one potential case of this like mentioned above).
Some people have suggested that there should be a community consultation phase built into the process where after learning about the issue at hand and putting together an initial draft of the legislation, the selected members return to their communities and host town halls or other events where the selected members get to explain their decisions and what they’ve learned to their communities and community members can offer feedback. I’m uncertain of how useful these would be in practice and the selected members would be in no way forced to follow the feedback they receive but perhaps it would help to have a built-in mechanism by which the selected members interact with their communities while they’re in the legislature.
Less Vulnerable to Corruption/Political Capture or More?
Once we understand that, usually, the goal of the selected members in sortition isn’t to try to choose what they think people in general think is best but rather what they themselves think is best, it’s easier to understand why some advocates of sortition actually suggest that the selected members’ votes should be anonymous. One benefit of this is that it makes the members harder to bribe for the same reason that making the voting of people in election-based representative systems anonymous helps prevent bribery. Additionally, many election-based systems require candidates to raise money to run for office and it may be easier to evaluate when there is quid quo pro between the representatives and the special interests in sortition than in those systems.
However, depending upon how the system is set up the biggest concerns about political capture may not be about the capture of selected members themselves. Many of the proponents of sortition also suggest having some way to have experts present to and advise the selected members when they are working on policy issues related to their field. This could be very helpful to inform the members about the issues they’re dealing with and help people without any prior experience crafting legislation do it for the first time but it also opens up the opportunity for the influence of special interests if they are able to influence the experts or how the experts are selected. There are a number of approaches to try to protect against this such as requiring experts to be accredited by international recognized institutions and having strong disclosure requirements. We could also have the expert selection process be conducted at least in part by an accredited community of experts (like what the American Bar Association does in the U.S. by giving ratings for proposed United States Supreme Court nominees). Certainly, none of these methods would be foolproof though. It is worth keeping in mind that expert capture is also going to be a concern in any election-based alternative but perhaps it would be an even bigger concern in a system where experts planned a more explicit role in the process and were dealing with novices.
No Standards for Competence, Interest, Values, etc.?
This leads to another one of the largest concerns that people have about sortition which is that there are absolutely no restrictions (or extremely mild restrictions) on who can be become a representative. In other words, people are worried about there being almost no minimum standards of intelligence, interest in the issues, moral character, *insert trait you value in your leaders here*, in order to obtain a significant amount of power. A key point here though is to consider what exactly do these standards look like, or what could they look like, in the election-based alternative. It certainly isn’t hard to come up with anecdotes of when the standards for who could gain power in election-based representative democracies weren’t nearly as high as some people would have hoped. However, I do certainly get where people are coming from with this concern. Currently, an important crux of this issue for me is how well the expert information sessions and advising could function. One thing that might make this process of trying to raise the competence of those selected easier is to have single-issue legislators. This could significantly lesson the burden of what the selected members are expected to learn about and understand and would probably be more feasible than in an election-based system where you could be dramatically increasing the number of candidates people have to know about. However, it certainly raises its own questions, like how do you properly partition up the policy space and how do these single-issue legislators coordinate with each other. Additionally, there may be some benefit to having representatives that have to think about things holistically.
Empowering Ordinary People?
One thing some people like about sortition deals with the flip-side of the previous concern. Another potential benefit over system where candidates have to raise money to run for office is that it greatly broadens who can be a political leader. This is because the selection procedure does not favor those who have pre-existing political connections, wealth, and fame which, although it could be affected by design choices like whether campaigns are privately funded, seems likely to be an issue in most election-based systems. On the other hand, some people worry that not having elections could decrease political engagement and result in an even less political informed and active populous.
Dis-Empowering Political Parties?
Once again, this is dimension that election-based representative systems already vary along considerably with, for example, voting methods like single-winner plurality-rule seeming to encourage a two-party system and penalizing independents more than others. However, a sortition system seems like it might lead to the loss of political parties all together. Currently at least this seems like a strong upside to sortition for me.
Not Viewed as Illegitimate?
Another concern about sortition is that the representatives may not be viewed as legitimate leaders since the general public didn’t play any role in selecting them. If people feel like elections are necessary for them to give consent to be governed over then they will feel like the randomly selected people don’t have the consent of the governed. For me personally, this way of thinking doesn’t hold too much sway since a system that will almost certainly do the same thing regardless of what you say hardly feels like allowing individuals to give consent in any way in which we would usually think about the concept. Additionally, for me, pretty much all related procedural concerns about sortition reduce to consequentialist concerns about how good the outcomes it leads to will be. But again that’s only how I personally think about the legitimacy of government and if others did view sortition as less legitimate and this led to viewing the laws it produced as less legitimate and more political instability then that seems like it’s worth taking into account. This gets at a larger issue which is that the performance of political and economic institutions is inevitably tied up in how social norms evolve around them which seem very hard to predict in general.
One point that I’ve been trying to make salient throughout this is that these two options are each a large category of institutions among which there may be a significant amount of variation in terms of performance based on specific design choices. I’ve also alluded to several choices that arise when designing a sortition legislature. Here are those ones as well as a few others:
- Who exactly is in the pool the candidates are selected from?
- Are there multiple legislative bodies and if so are certain bodies targeted towards specific issues?
- How long are terms and are members for a chamber selected all at once or in a staggered way?
- How well are the selected members compensated and what rules are in place to prevent them from being penalized for accepting the position? Are there any penalties for not accepting the position when selected?
- Is there some sort of community consultation phase or other mechanisms for the general population to offer feedback and be involved?
- Is voting on proposed legislation anonymous?
- What kind of disclosure requirements and ways of monitoring for corruption are there?
- Is there some sort of “learning phase” where members hear from experts in fields related to the policy issues they’re dealing with, if so how are these experts vetted and how involved in the process are they?
A Blending of Systems
So far I’ve been discussing things as a choice between a purely election-based and a purely random selection-based representative system but there are also a wide range of institutions in between (as well as ones that combine other elements such as those of direct democracies or more technocratic aspects). A simple way to combine these two systems would be to have some representatives randomly selected and others elected such as having a bicameral legislature like in the U.S. but with one chamber randomly selected and the other having elected officials. It could also be more specialized such as having the elected representatives create legislation and having the randomly selected members be responsible for voting on it or having single-issue randomly legislators like mentioned above and having the elected officials play a role in setting the agendas of and creating these single-issue groups.
An interesting variant of this kind of combination is to have the ratio of elected legislators to randomly selected legislators be directly determined by voter turnout. So for example if there was 65% voter turnout then 35% of the legislator would be assigned by random selection. In a sense then each absent voter is implicitly voting for a randomly selected person over the available candidates.
There are also a number of other ways to combine these systems such as randomly selecting the people eligible to be candidates in each election or only using sortition for certain kinds of decisions or only under certain political conditions (e.g. to resolve a legislative stalemate or institute certain political reforms).
While it can be hard to tell how institutions very different from what we can currently observe would work in practice, in some cases it may be possible to experiment with some of these alternatives on a small scale with ways to scale them up if they work well. Like I said, this is something I’ve only recently started considering so I’m interested to hear more points that I haven’t thought about and pin down more cruxes of this debate.