Ontario Citizens Assembly on the electoral system launched

At last…

A group of randomly selected Ontario residents will soon be asked to decide whether the province’s electoral system needs a shakeup.

The province said Monday it will convene a 103-member citizens’ assembly… If the group recommends a change, the government will then hold a referendum on the issue some time within its current mandate… Selection of panel members will begin this spring. [read more]

20 thoughts on “Ontario Citizens Assembly on the electoral system launched

  1. One interesting innovation with Ontario’s Citizens Assembly is that its legal terms of reference include eight principles it is to follow:

    An electoral system should have the confidence of the people and reflect their values.

    Fairness of Representation
    The Legislative Assembly should reflect the population in accordance with demographic representation, proportionality and representation by population, among other factors.

    Voter Choice
    An electoral system should promote voter choice in terms of quantity and quality of options available to voters.

    Effective Parties
    Political parties should be able to structure public debate, mobilize and engage the electorate, and develop policy alternatives.

    Stable and Effective Government
    An electoral system should contribute to continuity of government, and governments should be able to develop and implement their agendas and take decisive action when required.

    Effective Parliament
    The Legislative Assembly should include a government and opposition, and should be able to perform its parliamentary functions successfully.

    Stronger Voter Participation
    An electoral system should promote voter participation as well as engagement with the broader democratic process.

    Voters should be able to identify decision-makers and hold them to account for their actions.”

    Fair Vote Ontario Chair Dr. Joseph Murray said “In particular, we commend the government for noting that ‘fairness of representation’ and ‘proportionality’ — votes compared to seats in the Legislature — is one of the important principles in assessing voting systems.”

    The explicit reference to “demographic representation” is a victory of sorts for women’s groups who asked that fair representation for women be specified as a goal. Note that the government still doesn’t dare use the word “women”. Only half the population, after all.

  2. Somehow “fair representation” always seems to mean proportionality by gender/language/major ethnic group. Rarely does anybody demand that a certain fraction of representatives be immigrants, or poor people, or atheists, or nurses….

    I’m actually very conflicted over this issue, though. I’m uneasy with the idea of imposing restrictions on party lists, it seems to deny the basic principle of democracy to say “you can’t vote for this group since they’re all of the same ethnicity/gender/whatever”. Especially when taken to its potential conclusion, where parliament is pre-divided by demographics à la Liban. And if each voter is restricted to voting for someone from his or her group, that’s just completely wrong.

    On the other hand, it’s a real shande that our representative bodies are overwhelmingly made up of members of the dominant groups. This is one of the reasons why I’m very interested in open-list and flexible-list systems, it’s nice if there’s a way to say “I like this party’s policies, but I’d really like to be represented by a woman/Sikh/bilingual/whatever”.

  3. Vasi, your interest in group/factional representation within parties, and the multitude of dimensions on which it may vary, is precisely what my main research agenda is these days. I hope to make some headway on cracking this tough nut in the coming months (OK, years). My normative starting point is very similar to yours: Groups should not be pre-defined, and a party-list (or other) system ought to allow for the representation of groups whose interests cross-cut the parties to emerge as “naturally” as possible. It’s a complex subject that I hope to develop here in the near term, as well as in future published research.

    One qualification on Lebanon. As I understand it, the Lebanese electoral system would not be best characterized as one in which “you can’t vote for this group since they’re all of the same ethnicity” or other pre-defined group. It is true that the balance of sectarian groups in the legislature is set ahead of an election, but not because only people of a given group may vote for the candidates of a given group. In fact, quite the contrary. These legislators can be called inauthentic representatives of their group, because they may be elected by the votes of others. Each slate has to be balanced by the pre-set quotas, with that quota varying by district. But it is a general pool of votes, in that everyone in the district gets a ballot with the same set of slates and candidates.

    And then you have other systems with ethnic set-aside seats, like Colombia and New Zealand, but with very different rules. In Colombia, any voter may vote in either a nationwide ehtnic district (where the candidates must be recognized as members of the group, e.g. Afro-Colombian or indigenous) or in their regular “general” constituency for everyone else. So, the indigenous are guaranteed to have a certain minimum number of members, but those members may have been elected by non-indigenous voters (as indeed has been the case).

    In New Zealand, on the other hand, only Maori may vote (or run) for the special Maori districts, although they may opt instead to vote in a general constituency. I find both forms of group representation somehow distasteful, even though, in their own way, they are more “liberal” than Lebanon’s in the sense of providing more choice.

  4. Yes, what I meant about Lebanon was not that voters can’t vote for some of the existing slates. Rather, I meant that certain slates can’t form at all, because they don’t have the correct confessional make-up.

    Take for example an Islamist Shiite voter in Lebanon. Although he may wish to vote for an all-religious-Shiite list, he is not given that option. Instead he must vote for a slate that includes some Maronites (or Sunni, etc). So for anybody who prefers an all-Muslim list, it’s true that he “can’t vote for this group since they’re all of the same religion”.

    Thanks for the info about Colombia, I only knew about New Zealand (thanks to Jonathan) until now .

  5. In Lebanon, “Certain slates can’t form at all, because they don’t have the correct confessional make-up.”


    (And isn’t it interesting that we are talking about Lebanon–and Colombia and New Zealand–beneath a planting on Ontario!)

    (Oh, and I know a lot of things thanks to Jonathan. And thanks to you, too, Vasi!)

  6. I note you find forms of group representation somehow distasteful. I find them abnormal, but I must be wrong. The writers of the constitution of India, trained by elections in the British mode at the state level for generations, had no trouble creating reserved constituencies for Scheduled Castes and Scheduled Tribes, with a sunset provision; they expected the need for these transitional provisions to evaporate in ten years. They have been extended, extended, and extended some more.

    The world’s largest democracy can not only manage with 39 parties, and with a FPTP electoral system for 57 years, and with fine-tuned anti-party-hopping laws. It also has perpetually unsunsetted reserved constituencies, now extended municipally to be cross-reserved to 1/3 women, including a few reserved for ST women and SC women. There are electoral reformers who propose proportional representation in India, but I see no one proposing to abolish the reserved seat system. In fact some people want to extend it to reserve seats for OBCs (Other Backward Castes), not just for SCs and STs.

    Somehow it all works. If I don’t understand how, it’s likely for lack of trying.

  7. It is not group representation that I find distasteful. It is set-asides, or pre-defining what groups will be repersented that I find distasteful.

    “Somehow it all works.” Yes, I hear such things often about India, a truly amazing democracy, and apparently not one anyone would ever try to copy as a model. Yet, somehow how it all works!

  8. Sometimes territorial divisions are disguised group representation, as in Italy where there was a single member seat for Valle D’Aosta so that the French speaking minority could be guaranteed a seat, and I think similarly for the Sud-Tirol. In some Land elections in Germany minority parties are excused from thresholds (the Danes in Schleswig for example). Perhaps there isn’t a simple divide between systems with and without reserved seats. Would the Schleswig Danes, or the Valdotaine French, be justly aggrieved if their electoral arrangements were ‘normalised’? To some extent the same applies in the small Gaelic constituency of Western Isles in Scotland.

    One of the problems of setting representation for groups in divided societies is that it becomes very difficult and political to adjust representation. Lebanon is an example again – in the pre-civil war constitution taking a census became impossible because it would presumably have revealed Lebanon to have a Muslim majority.

    Prima facie, STV in large districts would seem to be the most ‘natural’ way to ensure group representation, but I realise that life is never that simple!

  9. Current gerrymander regimes —

    Half the votes in half the gerrymander districts = about 25 percent ANTI-Democracy minority rule.


    Total Votes / Total Seats = Equal votes needed for each seat winner — via vote transfers (using candidate rank order lists).

    Much too difficult for armies of math MORONS.

  10. In a few months Ontario’s Citizens’ Assembly will begin receiving submissions. They’ll need several detailed models on MMP systems to consider.
    My draft is almost ready for submission. Since this is the blog of the man who wrote the book on Mixed-Member Systems, it seems like a good place to ask for comments.

    Canada now has two citizen-designed MMP models. Both MMP models have
    the same basics: 60% local MPPs, 40% from regional lists; a province-wide “top-up” calculation; a province-wide threshold (Quebec: 5%, BC: 3%). Will Ontario’s be similar?

  11. Wilf, thanks for pointing out your draft, I like many of your ideas. Here are some comments–of course I’m not the expert here, so I’ll defer to Prof Shugart for the heavy analysis.

    I really like the multiple-selection flexible lists. Swedish-style flexible lists are fine for inserting a candidate into a party list–but they offer no practical way to reject a candidate from a list. With your system of multiple selections a voter could select everybody on a list except the unwanted candidate. This creates a sort of “negative vote”, for the party but against one unacceptable candidate. (Although this negative voting wouldn’t be practical in a large region, since you’d have to select twenty candidates!)

    It’s hard to be sure whether the flexible lists will be balanced properly, so that lists can move around somewhat but don’t display near-randomness. A lot of countries seem to have gotten this wrong, so my software engineer side suggests some “user testing”. Maybe draw up a sample ballot for your region that includes the names of real politicians–then give some random citizens a quick simplified explanation of how voting works under this system, and ask them to fill out the ballot. The results could be enlightening, maybe I’ll try it with some friends 🙂

    One thing that I’m slightly confused about is the initial selection of members from regional lists. You write, “First each region gets seats for the two most under-represented parties (highest remainder.)” My confusion is what happens when a party has very few or zero compensatory seats. In regions where they are one of the two most under-represented parties, do they still receive these seats even though it will cause or exacerbate an overhang situation? And if not, what’s the mechanism for deciding in which regions they will or will not get these seats?

    Finally, I think your draft could benefit from a more complete example. There are many steps involved in this system–I count about five–and I found that the overall picture got a bit lost by the end. An appendix going through every step from local seats to list placement in a minimal election might have helped. By “minimal”, I mean something like an imaginary province with two regions of eight and four seats, and very low vote counts–so it would hopefully not be too long or too much work.

    I hope you find this useful!

  12. What happens when a party has very few compensatory seats? The statement “in order of highest remainders until the party runs out of seats” was meant to apply to the first seats in regions as well, which I have now clarified. Thanks.

  13. New Brunswick is proceeding with an MMP model. It will go to a referendum May 12, 2008, which will require only 50% support. As Premier Bernard Lord rightly noted “Fifty percent has been the traditional number for democratic decisions.”

    “Mixed member proportional representation best meets the criteria for change that should define any new voting system in our province” says the government. First, alternative MMP versions will be considered and refinements made. Bernard Lord himself added “I’ve been in favour of this for quite some time. I like the concept of Mixed Member Proportional. Whether it will be the one developed by the Commission or a similar modified model that will be proposed to the public remains to be seen. We’re embracing the concept, the principles.”

  14. Pingback: Fruits and Votes

  15. Thanks to a posting by Wilf Day on another site, I just discovered this site. With the Ontario Assembly about to start its final deliberations, here’s a problem I’ve been looking at: How do you best assign the proportionally-allocated seats in a regional MMP system?

    Example: Assume a regional MMP system was in place in the 2003 Ontario election. Also assume a 2% provincial threshold. The Greens, with 2.8% of the provincial vote, would be entitled to about four seats in a 135 seat legislature. One regional model (Morrow) submitted to the Ontario Assembly uses a 9-seat region centred around Ottawa. In 2003 the Greens received 4.3% of the vote in this region – their highest regional percentage in Ontario. In the proposed 12-seat Southcentral region (including Guelph and Owen Sound), they received 3.8% of the vote. In the proposed 12-seat region of Eastern Ontario, they received 3.7% of the vote. In the proposed 27-seat Toronto region, they received 3.3% of the vote, and in the proposed 15-seat York-Durham region they received 2.9% of the vote. Using a system that assigns seats based on the highest fraction of a seat, the Ottawa region is at the bottom of the list with a .387 fraction, and likely would not get a seat under this assignment system. Yet, arguably, the Ottawa region is where the Greens have the most support. I would like to hear comments about this problem and any references that point to a relatively simple way of resolving it. Thanks.

  16. Intersting question, but out of date as to Morrow’s model which he has revised. As you’ll see, the four Green seats are in Toronto, Simcoe-Upper Grand-Huron, Peel-Halton and Ottawa-East. He details how to adjust any rounding discrepancies, and I agree. It’s not exactly how Sweden’s pure-list system works, yet it seems to me to be how a Swede would run an MMP model if they wanted to keep the number of MPPs from each region unchanged.

  17. An early consensus. The Citizens Assembly voted on Sunday on their first preferred alternative system. They plan to design two, and then choose one.

    Mixed Member Proportional – 78
    STV – 8
    Parallel – 6
    List PR – 3
    Alternative Vote (IRV) – 2
    Two Round System – 0

    That’s a lot stronger consensus than most expected.

    On Saturday they settled their three key objectives for system design, after breaking out into five group sessions. Chair George Thomson quipped “you’re making my life easy” when all five groups chose the same three:

    “The number of seats a party wins should closely reflect its vote share;”

    “Each MPP should represent a geographic area of the province;” and

    “Voters should be able to indicate their preferred party and candidate” separately, that is, have two votes, one for the party, one for the local candidate.

    The groups explained their reasons for proportionality: fairness, no wasted votes, each vote should go to electing a representative, and it “reflects Canadian values.” A teacher quoted his grade 9 student as saying “it just makes sense.”

    A PR-sceptic had “misgivings.” He wanted the objective to be “more closely reflect its vote share” but it said “closely.” The member from Temiskaming was fast with his rebuttal. “To say more (proportional) than we have now is ridiculous, ’cause, you know, throwing darts at a door is better than we have now, so let’s get real . . . just to say “more than we have now” is, you know, jeez, why are we all here.” (Applause.) Still, how closely is a decision for the design stage, as the Chair pointed out.

    Next step: preliminary design of the first alternative (MMP) on the weekend of March 3 and 4. Apparently that’s two-vote regional MMP.

  18. Pingback: Fruits and Votes &raquo

  19. Pingback: Fruits and Votes &raquo

Leave a Reply

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

WordPress.com Logo

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

Google photo

You are commenting using your Google 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.