Sortition Experiment

Debates on this forum and elsewhere lead me to conclude that there are, broadly speaking, three schools of the thought regarding the political potential of sortition:

1. The Blind Watchmaker

According to this school of thought, outlined in Oliver Dowlen’s Political Potential of Sortition and Peter Stone’s Luck of the Draw, sortition is primarily a mechanism to defend the institutions of government from corruption and partisan influences. Although historically associated with democracy there is no necessary connection as sortition could be applied to the selection of members of any group – democratic, oligarchic, aristocratic, associational or otherwise. Such an argument requires no empirical confirmation as it is true by definition (if it didn’t work then the process would not have been properly randomised). Chance (an arational process) precludes intelligent design, hence the (Dawkins) Blind Wachmaker allusion.

True believers, however, claim that sortition can also be used to produce representative democracy, but the claims here are divided into two camps:

2. Polytheists

Theorists in this camp, including Zakaras, Liebe, Sutherland and others, argue that sortition is primarily a decision-making process (deciding the outcome of a debate by secret vote), which should operate alongside the other institutions of liberal democracy, including election/refererenda and government executives appointed on merit.

3. Monotheists

Believers in the one true god of sortition would grant all powers, including the setting of legislative agendas and the appointment of government officers to the allotted body or bodies. The argument used in favour of allotment is that it is necessary in order to free politics from the domination of rich and powerful elites. Although advocates of this position, such as Callaghan & Philips, O’Leary and Gat, would advocate the incremental introduction of allotted assemblies (for example an allotted House of Representatives alongside an elected Senate and President) there is no conceptual division of function involved.

It strikes me that the claims of models 2 and 3 to provide democratic governance are open to empirical testing (not required for model 1). Fishkin (2010) claims that his deliberative microcosm ‘offers a picture of what everyone would think under good conditions. In theory if everyone deliberated, the conclusions would not be much different’ (2009, p.194). This is because a deliberative poll has a limited prescribed agenda and functions primarily to judge the outcome of a moderated debate (by secret voting), under conditions of balanced information and advocacy. If Fishkin’s claim is true then any number of concurrent deliberative polls on the same topic should have the same outcome. Although criticisms might, and have, been made of the DP methodology, nevertheless an empirical test of multiple concurrent DPs would indicate whether or not the DP procedure does or does not indicate ‘what everyone would think under good conditions’. Given the above methodological caveat (what do we mean by ‘good’) nevertheless if all (or most of the) concurrent debates came to the same conclusion it would not be unreasonable to claim that the DP performed its (limited) function in a manner that was fully compatible with democratic norms.

It would be equally possible at the end of each of the concurrent DPs for the moderators, advocates and information providers to withdraw and say to the allotted members, OK the DP is over, now its up to you to decide what your own political priorities are – what would you like to debate and what would be your legislative agenda? It’s plausible to imagine that each DP might not come to the same conclusion (unless there was a single issue dominating public debate at the time), as active functions such as agenda-setting and advocacy tend not to be equally within the comfort zone of all citizens – some feel stronger than others and there are marked differences in confidence, rhetorical ability and perceived status between different citizens. Without wishing to pre-empt the results, if it were the case that the experiment indicated a wide range of agenda-setting priorities then in order to make the process truly democratic a large number of allotted bodies would be required, followed by a further allotted body to decide between the various proposals offered. It would be hard to shield such bodies from the influence of lobbyists and it might therefore be better to leave the agenda-setting and advocacy role to other elements of a mixed constitution.

Such an experiment would be easy to conduct and no doubt the Stanford Institute would undertake it given the necessary funding.

11 Responses

  1. It is futile and obfuscating to pose the question of political decision as if it were a matter of determining what is objectively right or wrong, or even just a matter of getting a consensus. In any practical matter, within very broad limits of what we know to be true or false, it is a matter of constructing a choice that is partly a gamble on events in the future and partly a matter of what we think we want. Relying on any set of decision-makers can never be either epistemically nor morally conclusive, but if we are to act for common purposes we need decisions that are accepted by those who are affected by them, validated by conventions about how they are to be made. Those conventions need to be grounded in agreement about whom we can trust not to make their decisions on the basis of perverse or irrelevant considerations. Whether they will turn out to be optimal is going to be a matter of debate, even in the long run.


  2. What is the difference between “defend the institutions of government from corruption and partisan influences” and “free politics from the domination of rich and powerful elites”?


  3. John, my argument assumes that democracy is a non-negotiable premise, hence the need for experimental demonstration of what forms of sortition would or would not reflect “what everyone would think under good conditions”. The emphasis is on the word “everybody” and the experiments would indicate whether or not such agreement might be extendable from the judgment to the agenda-setting function. My hunch is no, but it’s only a hunch — hence the need for experimental verification. We disagree on the need for such a democratic mandate and on whether there is such a thing as the “right” answer. Epistemic democrats assume there is, and that all decisions can, ultimately, be resolved into a series of binary choices. Would I be right to conclude that demarchy would be best described as fitting into the Blind Watchmaker category in this post, in so far as a demarchic committee need not be democratic (the sortition could be made from any population, although you choose to privilege those with a self-declared interest). In your autobiography the examples of the committee structure that you favour (senior academics) are contrasted to the democratic rhetoric of 1970s student activists. I should add that I am something of a reluctant convert to democracy — my first book, as you know, made no allowance for voting or political parties — but I was persuaded of the need to make the proposal compatible with current mores. It seems to me that all those who are working within the subject area of deliberative democracy should be honest as to whether or not democracy is an essential part of their model, in most cases I suspect that it isn’t as “deliberative democracy” is something of an oxymoron.

    Yoram, defending the institutions of government from corruption and partisan influences is a mechanical process that works whatever the sortition base (plutocratic, aristocratic, democratic etc). Freeing politics from the domination of rich and powerful elites is an essential normative goal but there is no guarantee that it can be achieved by the simple application of sortition. This is because, as you acknowledge, all allotted members are only formally equal and because if the allottted assembly (or assemblies) had an agenda-setting role they would be open to corruption by rich and powerful elites. The way to ensure that politics is not dominated by rich and powerful elites is by careful checks and balances along with the adoption of innovations (such as the e-petition) that potentially empower ordinary citizens.

    But as the theme of this thread is a proposal for an experimental test to break this speculative log jam, perhaps we should focus on this instead.


  4. As usual, you use verbiage as a substitute to making sense. Domination by the rich and powerful is literally a specific case of corruption and partisan influence. Thus, if countering corruption and partisan influence were “a mechanical process” (which, of course, it is not), this would also be true about countering domination by the rich and powerful.


  5. If only it were that simple. All sortition does is ensure, by a mechanical process, that a target population is sampled randomly — thereby protecting the appointment process from corruption and partisan influence (as clearly demonstrated in Olly Dowlen’s book). However, once appointed, if the assembly were to set its own agenda allotted members would be an easy target for lobbying by the rich and powerful — and would not be protected by the binding mandate of manifesto commitments and/or party whips. If an MP belonging to a socialist party advocated, say, the privatisation of the BBC to Rupert Murdoch’s NewsCorp, then electors would punish that party in the following election, but would have no redress against a randomly-selected assembly member. No doubt you would seek to counter this sort of corruption via the criminal law but lobbyists are usually well ahead of the game in this respect — News International managed to illegally pay several million pounds to distant relatives of police officers in return for their services and got away with this for many years.

    This would not be the case if all allotted members did was to vote in secret, as lobbyists would have no way of ensuring that they got the necessary bang for their bucks. Just because an appointment process is a shield against domination by the rich and powerful, this does not insulate the resulting assembly from such influences. My additional concern is that an allotted chamber with active powers would also be undemocratic (for reasons that I’ve stated at length elsewhere), so it looks like a lose-lose situation.


  6. > protecting the appointment process from corruption and partisan influence [emphasis added]

    Now that does make sense. Where is the crucial clause (emphasized above) in your exposition? Without it, your exposition is senseless.

    Of course, I doubt that either Dowlen or Stone would be interested in sortition if its only effect would be to make delegate selection a process that allows no political manipulation. After all, there are many other automatic selection processes that allow no political manipulation. It is its function as a tool of democracy that makes sortition interesting.

    > once appointed, if the assembly were to set its own agenda allotted members would be an easy target for lobbying by the rich and powerful

    We have covered this ground many times before. Even if it were true that agenda setting is more vulnerable to corruption than voting (and I see no reason to believe it is), introducing a corrupt element (elites) explicitly into the process does nothing fix the problem. Your standard-issue rewards-based theory for the representativity of the electoral process does not bear scrutiny (see here). If it did work, then why would you need to resort to sortition at all, at any stage?

    As for the idea that voting in secret would shield the delegates from outside influence – this seems very unlikely to me. Secret voting is likely to reduce both the competence and the representativity of the delegates by reducing their motivation to carry out their job with appropriate thoughtfulness and diligence and by creating an atmosphere of unaccountability. It is also likely to produce immense skepticism among the public about the propriety of the process. (In general, the idea that secrecy reduces corruption contradicts common sense and daily experience.)


  7. Yoram,

    The point of secret voting is to make it difficult for would-be vote buyers to be sure they are getting anything for their investment in bribery (the corrupted representative can’t prove she delivered). This makes vote buying less likely. Other sorts of “corruption”… mere undue influence from expensive lobbying efforts, would not be solved by secret voting…That may require sequestration like for a jury, to prevent any contact by lobbyists.


  8. To be clear…lobbyists would need to go through standard channels to submit testimony, just like any citizen or interest groups, such that the amount of money a group had would not be relevant.


  9. Yoram, I’m sorry, I didn’t make it clear in the post that the Blind Watchmaker perspective is to do with the appointment mechanism only. Oliver Dowlen is emphatic that his interest in sortition has nothing to do with representativity, and goes to great length to justify this by claiming that the ancients had no notion of proportionality. He criticises writers like Callenbach and Philips for claiming that sortition is a mechanism for ensuring descriptive representation. As for Peter Stone, I’ll leave him to defend himself in this respect.

    >why would you need to resort to sortition at all, at any stage?

    On account of the rational ignorance problem. Sortition is probably the only way of reconciling deliberation with democratic norms.

    I agree with you (and J.S. Mill) that secret voting is something of a mixed blessing, for the reasons that you give. But, as Terry argues, it’s impossible to buy a vote when you don’t know how someone is going to use it [unless you buy all the votes] — this was the reason for the introduction of the secret ballot in parliamentary elections. I’d also concur with Terry’s strictures as to the standard channels that lobbyists would need to go through to submit testimony, and the possible need for sequestration. I’m afraid there is no way of abolishing the influence of elites by fiat, all you can do is to design institutions in such a way as to minimise their influence.


  10. > The point of secret voting is to make it difficult for would-be vote buyers to be sure they are getting anything for their investment in bribery (the corrupted representative can’t prove she delivered). This makes vote buying less likely.

    I understand the proposed mechanism. I think that not only is it unlikely to be realistic, but secrecy is likely to have the opposite effect. In fact, the entire “up or down decision only” setup would probably have similar effects – diminishing delegate motivation and responsibility.

    Be that as it may, the matter of voting in secret is independent of the main issue: the introduction of explicit elite influence into agenda setting makes no sense as a corruption reduction strategy.


  11. Yoram, at the risk of repetition, elites cannot be abolished by fiat, so it makes good sense to design political institutions carefully in order to better ensure that their influence is kept to a minimum. No doubt my own e-petition (which goes live on Thursday) will sink without trace as I am neither rich nor powerful, but you never know — given the relatively new phenomenon of viral marketing. And if I was a member of a small group motivated by a crusading zeal no doubt we could do a lot to garner the necessary 100,000 signatures without the need for money and/or power. Nineteenth- and twentieth-century anayses of political sociology are not automatically applicable to twentieth-century society.

    The worst possible scenario, however would be to replace democracy (power to the people) with klerotocracy (power to an arbitrary subset) as a result of a misunderstanding of the political potential of descriptive representation. Remember that the synonym for descriptive representation is statistical representation, and statistical laws only apply in aggregate. My own mathematical skills are pathetic compared to yours (or at least I would hope so, given your chosen profession) so I’m puzzled as to why you continue to deny (or evade) this simple conceptual point.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: