Criteria and two proposals for the use of sortition in politics

The Dutch organization has published a document (Dutch, German, Français, English) with criteria for the application of sortition in a political system and with two proposals for the use of sortition in the European political system.

The criteria deal with how the agenda is set, the sampling system, the size of the allotted chamber, its service term, its powers – advisory or binding and the potential for manipulation.

The two proposals are:

  • a “transitional” system in which the size of an allotted chamber is determined by the number of voters indicating support for this chamber during elections, and
  • a ‘European Citizens Jury’ – an allotted chamber set up as a review chamber next to the European parliament.

From the introduction:

According to historical sources, our political system was developed to protect the elite AGAINST democracy (sovereignty of the people). An “Electoral Aristocracy” was installed (18th century). Nevertheless, this can be seen as a positive evolution compared with monarchy.

Later on, some “democratic” elements were introduced, for instance “free”, or so called “democratic”, elections with universal suffrage, the equality principle, freedom of speech, freedom of organization, free press etc. However, some of those elements were moderated or abolished afterwards.

But a “democratic element” is not yet a “democracy”. Freedom of organization may be a “democratic element” without which a democracy cannot exist, but on its own it is no democracy. Hence “free elections”, to appoint a governor for instance, may be a democratic element, but on their own they are by no means a democracy. Furthermore, our political system of representation by elected representatives originates from the Roman Republican system and not from the Athenian Democracy. Calling our political system a “democracy” is deliberately misleading propaganda.

The rise of political parties, which have in fact taken over legislative power, enabled lobbyists to work more directly. Thus, power concentration of legislative power and economic and financial interests was achieved, eliminating the last remains of democracy in our political system. One could say that history is repeating itself. The Roman Empire collapsed due to internal decadence and the emergence of a few wealthy and powerful people – this is apparently inherent to the electoral representative system.

Some countries, as well as many cities and communities, have already made the choice for a more democratic political system more than a century ago, by introducing the binding referendum with the citizens’ initiative. Other countries made this choice more recently. The best known examples are Switzerland and half of the states in the US.

But here we notice an evolution towards (or a demand for) further democratization as well, such as with the introduction of a sortition-based parliament. The main reason of this aim is the ever increasing power of the financial and economic interests and their influence on the legislative power. The system of sortition is based on the statistical fact that a sample, which was selected in a scientific manner, gives a faithful image of the whole population. As a result, decisions made by this group are representative for the whole population.

20 Responses

  1. It is written in page 8: “No mixed panels of professional (i.e. politicians) participants and people appointed by sortition”. Why not? I’ve never got a satisfactory justification.


  2. While individual cases of mixed bodies may happen to work, there are serious likely problems in terms of the psychology (leadership -followership – status – domination – cognitive free-riding – etc.) A deliberative body should be made up of formally equal members to allow free deliberation, and there is a marked danger of inequality in which some members are seen as “more equal than others.”


  3. In most cases that I know of mixed panels are a proposal of politicians or people who probably want to maintain the power of decision making where it is for the moment, with politicians, political parties and lobbyists with only one objective: maintaining the status quo.
    What is the most important property of a ‘political system’? It is trust.
    In the past decennia, trust in the electoral representative political system is going down (Edelman trust barometer) so it is normal that those who are involved in the electoral system will try to improve this with all kind of minor adjustments that changes the least possible in the execution of power. An additional possibility is that it becomes possible to discredit the sortition system by showing that it doesn’t work as thought or beneficial results are marginal.
    Nevertheless it will be difficult to prove that mixed panels are not at all a ‘democratic instrument’, but it is not impossible.

    We have the example of the Washington state mixed panel that decides about the salaries of elected officials. ( Citizens’ Commission on Salaries for Elected Officials )

     The Commission consists of 17 unpaid, citizen members selected by two methods:
    Ten members are randomly selected by the Secretary of State from the rolls of registered voters, one from each Congressional District.
    Seven members are selected jointly by the President of the Senate and the Speaker of the House of Representatives; one each from:
    private institutions of higher education;
    professional personnel management;
    the law;
    organized labor; and two members are recommended to the President of the Senate and to the Speaker, one by the chair of the Washington State Personnel Resources Board and the other by the presidents of the state’s four-year institutions of higher education.

    I presume that this example is a worst case scenario, for me at least. 10 citizens appointed by sortition can’t be ‘representative’ by any standard, even when scientifically stratified under ‘neutral’ supervision. And this selected panel is then mixed with all kind of ‘professionals’.

    It would be interesting to compare the results of this panel, when kept secret, with the results of a panel that is descriptive representative and decides after hearing the same experts and politicians. But I suppose this is up to the citizens of Washington.


  4. > 10 citizens appointed by sortition can’t be ‘representative’ by any standard

    Such statements are often made on this forum. In fact, this is far from clear. In the context of setting salaries, 10 people seems acceptable. 20 may be better, but 50 would quite probably be worse, since it would result in spreading decision making power too thinly.


  5. Yoram: >Such statements are often made on this forum. In fact, this is far from clear. In the context of setting salaries, 10 people seems acceptable.

    And wide open to corruption.


  6. I can understand that with a scientific stratification they can get a panel that reflects some ‘diversity’. Lets try (my own guess :-) ) One male, one female, one holebi, one old, one young, one poor, one rich, one living in the city, one from agricultural area, one homeless. In such cases the selected people also have to respond to questionnaires in order to avoid people with a clear point of view about the matter at hand. Even when we accept that all this ‘manipulation’ (called selection) is done with the best intensions in a scientific and neutral way, under supervision, (ref. CDC Jaques Testart in France) there is no way that such a panel is also proportional in his diversity.
    We only can notice that in a state where the time to do the job (Grey) is estimated the same, Nebraska, where salaries of elected officials is subject to referendum is 12.000$ a year and in Washington 45.474$ a year. Arizona is also a ‘gray’ state where salaries are subject to referendum and they have 24.00$ a year.


  7. This is an issue of an overlap of two distinctly different functions of sortition. We have statistical representativity on one hand (which requires far more than ten members, since there is a substantial probability of selecting an extremely UNrepresentative group); and on the other hand impartiality (nobody was able to hand pick people who would favor a particular proposal). For SOME purposes all we need is impartiality and sufficient common sense. In the case of legislative salaries it seems unlikely that accurate representation of most demographic traits will matter that much.


  8. In some cases a descriptive representation is indeed not needed. We can refer to the juridical application for this. I have a reference on this in the paper although I dit not study it in dept. Nevertheless I found that the initial main reason to install a citizen jury was the short time the jury was in charge. It would be interesting to compare the results of a descriptive representative jury, not mixed, with the actual design in Washington state. The main reason is the big difference with the referendum decisions. But there may be a lot of other reasons for this difference of course, except from the workload who is estimated the same.


  9. Terry:> nobody was able to hand pick people who would favor a particular proposal.

    I believe the figure of 10 persons has been suggested as the ideal size for a jury to set salaries. How do you propose addressing the problem of ex post corruption?


  10. > which requires far more than ten members, since there is a substantial probability of selecting an extremely UNrepresentative group

    Again, I don’t think this is obvious at all. It really depends on how you define unrepresentative (and to some extent what the decision making mechanism is).


  11. Yoram:> It really depends on how you define unrepresentative

    Terry was explicitly referring to statistical representativity, which would be impossible with a group of 10 randomly-selected persons (in anything other than an extremely crude sense). If that were not the case opinion pollsters could save themselves a lot of money by reducing the size of their samples. You have often claimed that your sortition model is based on statistical representation but you have never really explained what you mean by this (and it is doubly puzzling if you claim that a randomly-selected group of 10 would be statistically representative of the target population).


  12. > referring to statistical representativity which would be impossible with a group of 10 randomly-selected persons

    This of course is so meaningless, it is not even wrong. (Rather typical for you, unfortunately.)


  13. Please explain (for the sake of mathematical ignoramuses like myself and Terry).


  14. What Keith and I are getting at is, for example, suppose an ethnic or racial minority constitutes 30% of the population. If groups of just ten people are randomly drawn, now and again the sample will have NO members of that minority, or a majority of members of the group will be from that minority. If a sample of 400 are drawn instead the likelihood of such extreme unrepresentativeness drops close to zero.


  15. Terry,

    Agreed. Of course we can’t know the relevance of ethnicity, gender or any other population parameter wrt to a salary-setting jury (although socio-economic status is almost certainly a determining factor). Chances are that different samples of 10 persons would return different decisions, so the decision would not represent the considered view of the target population (irrespective of any epistemic considerations). This is what I think we both mean by “statistical” representation, so it would be good to hear from Yoram in what respect this understanding is “so meaningless, it is not even wrong”.


  16. Terry,

    On the issue of the salaries of officials, the presence of a particular ethnic minority in the sample is of no particular importance. What is needed is a reasonably accurate representation of the ideas relating to the appropriate level of salaries. Again, depending on how this notion is defined exactly, and how the body makes its decisions, a 10 people sample may very well be a “representative sample” on this issue.


  17. Yoram:> 10 people sample may very well be a “representative sample” on this issue.

    On the other hand it might not — what one can be sure of is that different samples would contain people with very different ideas and the LLN certainly doesn’t apply to a group of 10. The number is lower than that required for a trial jury and all the latter has to do is to rule on the facts of the matter by evaluating two competing arguments (leaving aside their own ideas). This being the case the sort of body that Yoram has in mind should not be described as a jury — a term which is tossed aroumnd promiscuously on this forum.

    And Yoram still has to tell us exactly what he means by “statistical” representation, as I’ve genuinely no idea if he claims it can refer to a randomly-selected group of 10 persons. All that one can say about such a group is that they are all likely to be human beings (perhaps that’s all that matters?), as even a crude population parameter like gender would not be accurately sampled.


  18. In the case under consideration (setting salaries) I lean towards the idea that accurate statistical representation is not crucial, and that impartiality is the goal. Ten people might be fine. There is no “correct” answer, nor is there even one answer that “society as a whole would make if well informed,” etc. Even one individual person might change her decision depending on the day … one day thinking the median salary in society is appropriate, and on another day thinking a higher salary might protect better against bribery, etc. The goal is to simply assure that those who will be getting the salary are not in control, and that the decision makers are not systematically biased in favor or against those who will be paid. And then, based on real world experience, a subsequent jury might change the salary at the next round to address problems of the salary level set by the previous jury.


  19. If impartiality is the goal in this case then using a ‘mixed’ panel is very questionable. And that was the original question I think. Nevertheless the use of 10 (Washington), 40 (Oregon CIR), 100 (Climate panel luxembourg), 1000 (G1000 Belgium), and so on, all with a very weak explanation and subject of criticism and doubt is in my view a weak spot for the sortition movement that needs to be addressed. For me each choice has to have a strong defence and a clear explanation, that can only be in our (proponents of democracy) benefit in the long term. I still try to find arguments and proof for the use of small panels and in what cases (I start to study the CDC of Jaques Testart) . If I can’t find a good argumentation and defence I stay with the ‘descriptive representative panels’. It is of no use to propose fundamental changes if you don’t have a strong case. That is why we developed the criteria in the first place. We not only have to make strong proposals, we also have to refuse implementations that will harm the sortition idea in the long term. How to convince people if we ourselves don’t believe in what we propose.

    Liked by 1 person

  20. I would be happy with impartiality (assuming some defence against ex-post corruption), but my interest is in Yoram clarifying what he means by “statistical” representation, as he appears to be using the term in a different way to everyone else on this forum, where Hanna Pitkin’s definition (of “descriptive” representation) is the accepted one. I think this is important, as it might help prevent us talking (or shouting) past each other in the future. As it stands both Terry and myself stand accused of a usage which is “so meaningless, it is not even wrong”. If we are going to make any progress in the study of sortition then we need to agree a common use of language.


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 )

Google+ photo

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

Connecting to %s

%d bloggers like this: