Accountability Under Sortition
From "The Trouble With Elections: Everything We Thought We Knew About Democracy is Wrong," Chapter 13.9
For thousands of years democracy through sortition (the use of lottery selection with rotation and political equality) was recognized to be a system for holding power to account. The ancient Greek historian, Herodotus, put this sentiment in the mouth of Otanes in the constitutional debates of Persia in his Histories. According to Herodotus, Otanes argued that the concentration of power, such as in a monarch, causes corruption. Greek democracy (by the name of equality – isonomia) was the best system because it is free of the abuses typical of monarchs, power is distributed widely by use of the lottery, so all officials and power are held accountable by the multitude.
The concept of accountability for corruption can be applied at the level of the individual member for an elected or lottery-selected law-making body. But, the very concept of representational (policy) accountability for an individual member (as it is applied to an individual elected representative1), is simply not applicable to a system of sortition. A mini-public as a whole is representative of the population as a whole. No individual member has any constituency. It can be argued that the term “a representative” isn’t even an appropriate term for a member of a mini-public. Elected representatives who deviate from the interests of their constituents can theoretically risk removal. Under a sortition system of randomly selected mini-publics, since no member has a constituency to be accountable to, each member is expected to vote in whatever way he or she thinks best, based on the information presented. In terms of policy accountability it is necessarily the body as a whole, rather than any individual member that is the appropriate level for scrutiny.
Random selection maximizes the sort of “natural” or congruent accountability described by Mansbridge as the alternative to sanction-based accountability. This is because the members are a microcosm of the public. Using John Adams’s maxim again: the mini-public is
“in miniature an exact portrait of the people at large. It should think, feel, reason and act like them.”
If presented with a draft law, any statistically representative group, or the population as a whole (if all citizens could somehow be engaged in the same deliberation) would be likely to vote in a similar way.2
The political theorists Cristina Lafont and Nadia Urbinati have argued that a sortition body can either be genuinely like the population, mirroring their preferences in a populist way, or become well-informed, serving as a technocratic filter for the poor judgment of the less-informed public, but that these two concepts are “mutually incompatible.” A sortition body cannot have the same preferences as the people, and also have better preferences.
“Both claims cannot be simultaneously true. Lottocrats must choose between these technocratic and populist claims.”
This is a rather silly objection. The (fully compatible) goals are to both select an accurately representative sample of the population, who share the same values and interests as the population, and to have such representative groups become adequately informed so as to make the same decisions that the population as a whole would make if they could somehow all become well-informed, hear opposing perspectives and analysis, and deliberate, without distorting personal power agendas – representative and informed.
As a thought experiment, imagine two voters with identical policy preferences. In this example, let’s say they are pro-union, anti-immigrant (viewing it as a ploy by business owners to bring in low-wage workers to suppress wages). They are both pro-life (anti-abortion) and view climate change as a crucial issue requiring urgent action. There is no political party or candidate that agrees with them on all of these issues. Perhaps one of them has a parent who always stressed how much the union improved their life, and convinces that voter to vote for the pro-union party on the left. The other voter saw a frightening advertisement about dangerous immigrants, and decides to vote for the party on the right. Neither has taken the time to become well-informed. Even though their policy preferences are identical, unpredictable circumstances prompt them to vote in opposite ways.
For most of the hundreds of issues, voters have no, or only a sketchy preference. If interrogated closely both of these hypothetical voters might say that they recognize that their policy preferences are not deeply thought through, and are just ideas they heard that made sense to them at some point in the past, but superficially. Their considered preference, once better informed, might evolve, or even flip. They might much prefer that people like them (without hidden agendas), were given the time, resources and motivation to really dig into each of these issues (and others that haven’t occurred to them) in order to make a truly informed decision on each. “I don't want my off-the-top-of-the-head answer on some opinion poll to be controlling. That is not accountability.” We should want our representatives to be willing to make an unpopular decision.
To add an additional layer of policy accountability, we may want to add a double check in the case of close votes. If the majority threshold of a jury vote is less than the acceptable confidence level for the sample size, in order to provide accountability and assurance of representativeness, a second jury could be called (perhaps larger) to see if they come to the same conclusion.
With less power (dealing with far fewer issues and no re-election longevity) sortition representatives would have less opportunity for corruption. The power that accrues to a politician who is a committee chair with lots of seniority and repeated re-election campaigns is a recipe for corruption. Such concentration of power maximizes the push and pull of corruption. It prompts interests to seek to establish a corrupt relationship, and multiplies the power-holders opportunities for corruption. However, a decision by a mini-public could be the result of corruption.
The prevention of corruption was one of the primary purposes of the lottery system in classical Athens. Athens had a system of rendering of accounts, scrutinizing the performance and financial honesty of all of its randomly selected magistrates and councilors at the conclusion of their service by randomly selected panels of inspectors. In a modern context, the same anti-bribery provisions that protect against jury tampering can be applied to sortition bodies. The threat of prosecution for wrong-doing should also remain. It would be important to protect against pay-offs to former mini-public members in cash, job offers, or other perks, from groups interested in the legislation the representative had worked on. All mini-public members might be told that an “undercover” person might pretend to try to bribe them, and if they report it they would get a reward.
If, however, a sortition system is poorly designed (by replicating some of the corrupting aspects of existing elected legislatures), the representational accountability could be hampered. Simply substituting selection by election with selection by lot is not enough. If a law-making body (selected by lot), dealt with a wide range of issues, and had a long term of office, it is likely that power hierarchies would develop, vote trading and other untoward practices would arise. The longer a given group is in power the less they would be typical of, or be representative of the population from which they had been drawn, and the more likely corruption would develop. This is one reason I advocate separating out the preparation process for proposals (where more extended experience is particularly helpful), from the final decision-making function, which ideally should be conducted by short-duration juries that deal with a single issue to prevent vote trading and developing a sense of power and entitlement typical of elected legislators. Rotation as practiced in Athens and many other jurisdictions is a key component for avoiding all kinds of corruption and unrepresentativeness.
Short duration decision-making legislative juries also have the beneficial proclivity to “regress to the mean.” That is, while one random jury may be a bit less representative than average, any unrepresentative decisions would tend to be modified by subsequent random bodies. This process of reconsideration of past bad decisions is enhanced, compared to long-duration elected legislators, simply because there is no pride of authorship causing representatives to cling to failed policies long after their merit has been disproved. With constant rotation and new mini-publics, sortition is a self-correcting system. A bad decision by one jury will likely be reversed by a subsequent jury. Unlike politicians who often double down on their previous decisions, each jury is fresh, without politicians’ motivation to save face by insisting that their earlier decision was right.
Accountability of the group as a whole is also improved by providing separate review and rules-making bodies. There is always a risk of corruption when a group with power defines its own rules – such as Congress setting its own election laws, ethics enforcement procedures, etc. The principle of separating powers, which has widely been recognized as a cornerstone for protection against undue concentration, resulting in tyranny, should be applied to rule-making for legislative bodies as well. In a nutshell, the person or entity that defines the choices or decision-making process should not be the same as the one making the final decision. In Federalist Paper No. 10, Madison wrote:
“No man is allowed to be a judge in his own cause, because his interest would certainly bias his judgment, and, not improbably, corrupt his integrity. With equal, nay with greater reason, a body of men are unfit to be both judges and parties at the same time.”
This basic principle of separating tasks was noted by the seventeenth century political philosopher and supporter of Cromwell, James Harrington. In his magnum opus describing a Utopian constitution, Commonwealth of Oceana (which also incorporated some sortition within its design), Harrington wrote:
“[Two children] have a cake yet undivided, which was given between them: that each of them therefore might have that which is due, ‘Divide’, says one to the other, ‘and I will choose; or let me divide, and you shall choose.’ If this be but once agreed upon, it is enough; for the divident, dividing unequally, loses, in regard that the other takes the better half. Wherefore she divides equally, and so both have right.”
Through the separating of tasks the children provide accountability to each other. This anti-corruption accountability feature is one of the reasons undergirding my multi-body sortition design. Elections almost universally fail to provide actual accountability, while well-designed sortition has accountability built-in by harmonizing the nature and diversity of the decision-makers with the nature and diversity of the population.
Some countries that vote for political parties rather than individual candidates may be seen to be holding an entire party to account through electoral removal. These representatives are more firmly locked into supporting their party line. But in the United States, accountability is expected to be achieved one candidate at a time.
Note that with an infinite number of variations of possible draft laws, it is not expected that different random groups would draft identical laws, This is one of the reasons I separate the drafting and approval of laws into separate bodies. While active deliberation by different random groups would develop different draft proposals, the tendency would be for each of them to have a reasonable chance of meeting with jury approval. This would not be the case with proposals developed by political parties or special interests.
👏
I thought this post covered some of the accountability risks of citizens assemblies, but it felt like it pulled some punches.
The author discusses 3 risks and potential solutions, given below for easy recall:
1. policy accountability risk: “If the majority threshold of a jury vote is less than the acceptable confidence level for the sample size, in order to provide accountability and assurance of representativeness, a second jury could be called (perhaps larger) to see if they come to the same conclusion.”
2. Bribery risks: “In a modern context, the same anti-bribery provisions that protect against jury tampering can be applied to sortition bodies. The threat of prosecution for wrong-doing should also remain. It would be important to protect against pay-offs to former mini-public members in cash, job offers, or other perks, from groups interested in the legislation the representative had worked on. All mini-public members might be told that an “undercover” person might pretend to try to bribe them, and if they report it they would get a reward.”
3. Representational accountability risks: “If, however, a sortition system is poorly designed (by replicating some of the corrupting aspects of existing elected legislatures), the representational accountability could be hampered. Simply substituting selection by election with selection by lot is not enough. If a law-making body (selected by lot), dealt with a wide range of issues, and had a long term of office, it is likely that power hierarchies would develop, vote trading and other untoward practices would arise…Rotation as practiced in Athens and many other jurisdictions is a key component for avoiding all kinds of corruption and unrepresentativeness.”
But the amount of words spent on these actual risks is only a fraction of the post’s total. Most of the words in this post are discussing the anti-corruption features of sortition. The previous 2 posts were a good place for these anti-corruption ideas, but this post should try to steelman the argument for why sortition is risky, and how to ameliorate those risks.
Also, there were some more risks I am surprised he didn’t discuss:
1. Policy accountability risk: What prevents CA members from voting in their short term interest, since they don’t have any external long term electoral pressures like politicians do?
2. Problems with the randomness of sortition: how can the public reliably verify that a sortition was done using true randomness? What are the ways that the random sortition process could be subverted, and how do we protect against it?
3a. CA members inability to write legislation: randomly selected citizens likely aren’t going to be lawyers, so they aren’t going to be good at writing laws. How do citizens assemblies reliably turn the representative citizen views into wise legislation?
3b. If the CA members aren’t writing the legislation, but instead hand off their output to a group of lawyers, how are the resulting principal-agent risks dealt with?
Overall, I think this post started to get at the risks of sortition, but it would be improved by removing some of the pro sortition paragraphs and replace them with accounts of additional risks to sortition.