Icelandic constitutional assembly–elected by STV

Iceland will have a Constitutional Assembly “for the purpose of reviewing the Constitution of the Republic.” It will be elected by single transferable vote (STV), in a single national district with a magnitude of 25. The election will be 27 November.

The STV system will include a gender-representation provision:

There is a 40% sex-balance rule that is to be applied after the STV-PR election for the 25 places has been completed. If candidates of either sex do not have at least 10 of the 25 seats (“two-fifths”), additional seats will be allocated to the under-represented sex to bring representation of that sex up to “two-fifths”, subject to a limit of six additional seats. These additional seats would be allocated to the required number of the last eliminated candidates of the under-represented sex.

See more at the Irish Political Reform blog. And thanks to Tom Round for pointing this out in another thread.

Iceland’s M=25 is likely a record for size of an STV constituency.

8 thoughts on “Icelandic constitutional assembly–elected by STV

  1. Setting aside the merits of constraints, gender or otherwise, the approach described here seems at first glance unfortunate. To begin with, STV’s order of elimination (or order of election, for that matter) has no intrinsic significance. Worse, the ballots supporting these last-eliminated candidates likely will have been transferred to candidates who are eventually elected, so the procedure works against proportionality; those voters double-dip.

  2. They could have achieved the same result without overhang seats by simply limiting each sex to to a maximum of 15 delegates.

    Query for STV junkies: I believe Enid Lakeman argued that 21 is the ideal STV magnitude. Anyone got a reference?

  3. In How Democracies Vote, the chapter on STV, Lakeman discusses the tradeoff of voter knowledge vs district size, mentions that Cork used a magnitude of 21 for “many years” successfully, and mentions in passing that there are diminishing returns in growing from 21 to 23.

    She doesn’t really argue for an ideal magnitude, though, at least not here.

  4. As the author of the note you quote, I must correct your statement
    “The STV system will include a gender-representation provision”. As the wording you go on to quote makes very clear, the sex-balance provision is NOT part of the STV system. That provision will be applied only AFTER the result of the STV election is known.

    It could have been done differently, “STV with constraints”, when the sex-balance requirement would have been built into the counting rules. But in Iceland they chose to go about it in another way.

    Jonathan Lundell is right to point out that the order of election in an STV-PR has no intrinsic significance. Other procedures must be applied if you want to order the elected candidates. See, for example:
    But that was considered a complication too far – on this occasion.

    As to Enid Lakeman reportedly arguing for 21 as the “ideal” district magnitude for STV-PR, I cannot supply a reference, but suspect she would be referring to the law of diminishing returns as applied to voter representation. See:

    Click to access Droop-Quota-Diminishing-Returns.pdf


  5. The justification for 21, or any other single number, requires more than the principle of diminishing returns. You also have to decide what it is about the shape of the curve that makes some point on it “optimal”.

    I think there might be a loose relationship between the “optimal” number and the size of the jurisdiction, and/or the effective number of parties/slates/factions. But I don’t know how to quantify that either.

  6. Lakeman: “…the advantage of extra members falls off very rapidly after the first few, and must soon be overtaken by the drawback of growing impersonality and consequent lower accuracy of representation of groups other than the parties. Constituencies each returning about five member would seem likely to give satisfaction in both respects.”

    OTOH, in the example of Cork (1950), she cites an 11% non-transferable rate as evidence that a 21-member constituency can be workable.

    My sense is that 5-7 is a good range, and that an example like Cork is likely to be the result of high party identification.

  7. Some statistics from the full official results:
    There were 522 candidates and 82 335 voters who could give up to 25 preferences. The ballot contained 25 boxes, one for each preference, and voting was by filling in the code number of the candidate. (For details see this how to vote book with all candidates)

    Only one candidate (3403 Þorvaldur Gylfason – M) was elected in the first round (7192 first preference votes or 9%)
    Ten others were elected by reaching a full quota (3167 votes) in counting rounds 306, 344, 425, 435, 469, 474, 479, 493, 500 and 506.

    The result was known after 508 rounds with fifteen candidates still competing for fourteen vacancies. (I see no reason to do the 509th round in which the votes of the last excluded candidate were redistributed over the fourteen winners.)
    In the 508th round, the fourteen last winners had 32 587 votes (ranging from 3047 to 1952 votes) and the last loser 1930 votes, while 12 981 were non transferable. So, 14 911 votes (18%) were ‘lost’ (1930 + 12 981) meaning they did not end up in the ‘pile’ of one of the twenty-five elected candidates, while 67 424 votes (82%) were ‘effective’ (11 quotas + 32 587).
    Of all voters, 31 673 (38%) gave their first preference to one of the twenty-five winners.

    There is a meaningful difference between the top-25 of the first round (only first preferences) and the twenty-five winners: after the first round, the eventual winners were ranked 1 to 12, 15, 16, 18, 19, 20, 22, 27, 29, 32, 33, 37, 39 and 40; seven of the top-25 of the first round (places 13, 14, 17, 21, 23, 24 and 25) were overtaken by lower candidates in the end (places 27, 29, 32, 33, 37, 39 and 40).

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 )

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.