I agree completely. Some of my earlier work in favour of IRO/ STV is online [see http://www.cs.mu.oz.au/~lee/prsa/res.html -- although I've recanted a few minor points in the 10-12 years since I wrote these, mainly that I'm now much more favourable to Condorcet systems for referenda and single-seat contests).

However, I've also found that many people who regard IRO with suspicion and mistrust have no such objections to runoff and/or primary elections, even though these ensure a roughly similar result to IRO, albeit with greater cost in time and money. Some of the shaky arguments used against Alaska's recently defeated IRO initiative ("IRO violates one-person, one-vote by giving minor-party supporters two votes"), for example, would have been laughed out of court if they'd been said about runoff systems, which are simpler and easier to explain to the average voter.

At 08:50 26-11-2002 -0600, Steven Mulroy wrote:

I think the terminological point makes sense, and the proposal is interesting.
But let me ask you this: wouldn't many of the concerns the proposal is trying to
address be taken care of by an "instant primary" or IRV system for single-member
offices, and a "choice" or "preference" rank-order system, using something like
the Hare system counting method, for multi-member legislative offices? These would
have the advantage of needing only a single election event. SJM

"Tom Round (home)" wrote:

> Perhaps this is a chance for political scientists to steer the definitions
> in a more precise and stipulative definition. How about something like ...
>
> 1.      A "primary" is an election where, within each party, the winning
> candidate[s] are deemed nominated or endorsed by that party for the general
> election. Ie, a later election must follow even if, say, 6 out of 8 million
> citizens vote in the Republocrat Party primary and of them, 5 million
> prefer Bloggs. (The only ground for dispensing with the later election
> would be if Bloggs were the only candidate: I assume the Supreme Court's
> reading of the "uniform date" federal provisions impliedly allows for this
> -- unlike, eg, Austria and Thailand, where even if s/he unopposed a
> candidate must face a yes/ no referendum.)
>
> It is therefore NOT a "primary" if a candidate can claim victory over one
> or more surviving rival candidates by gaining the votes of some prescribed
> quorum, majority or super-majority of the voters (or of the total enrolled
> electorate, or the Voting-Age Population: Peru, I think, used to require
> that the plurality Presidential candidate be supported by 33% of the total
> adult population, whether or not s/he polled more than half the votes).
>
> This usage would reflect the fact that "primary" is only used, outside the
> USA, for mass ballots limited to party supporters (who usually must be
> paid-up members) -- eg, in Israel to choose the Labour and Likud candidates
> for Prime Minister and to rank the party's Knesset list. The numerous
> countries that use runoff systems for the President and/or legislature do
> not seem to use "primary" (or its equivalent) for the first round.
>
> 2.      I propose "first round" or "preliminary ballot" in cases where a
> sufficiently popular candidate can defeat his/her rivals by outpolling them
> decisively enough on that ballot.
>
> After all, it is possible to have both a primary and a general election AND
> to have two separate rounds of voting for each, if no candidate gets a
> majority on either -- a total of four possible rounds.
>
> One possible model might be to prescribe a minimum of two and a maximum of
> three ballots for each office. On the first ballot (the "primary
> election"), any candidate can enter. Candidates who are registered members
> of the same party, and seeking the same office, are grouped together. A
> voter may select one candidate regardless of party.
>
> Then, only the following candidates may continue to the second round (the
> "general election"), a month or two later (I'll leave aside whether
> write-ins should be permitted):
>
> [a]     Either of the two candidates with the most votes (regardless of party)
> (skipping over any who voluntarily withdraw).
>
> [b]      Any other candidate who individually receives more than [say] 10% of
> the votes. (Modelled on the French rule, whose details fluctuate but which
> now, I believe, requires 12.5% of the total enrolled electorate for third,
> fourth, etc candidates to avoid elimination -- but not in Presidential
> elections where only the top two remain, as M Jospin found to his chagrin a
> few months ago).
>
> [c]     For each party whose candidates in total poll more than [say] 10% of
> the votes -- whichever of those candidates receives the most votes
> (skipping over any who voluntarily withdraw).. (So five Greens could run
> against a single "unopposed" Democrat without losing the Green slot on the
> general election ballot, provided that at least 10% vote Green in the
> aggregate.)
>
> Finally, if no candidate received 50% (or 40%, or whatever) of the votes in
> the second "general election" round, a third, "runoff election" would be
> held, a fortnight later. Only the two highest-polling candidates from the
> second round may remain (skipping over any who voluntarily withdraw).
>
> I'm floating this thought-experiment as a question of good policy, without
> delving into its constitutionality -- especially the question of whether
> associative rights would be unconstitutionally infringed if the XYZ Party's
> official nominee could be selected in any part by voters who have
> acknowledged no allegiance to the XYZ Party, even by the fairly minimal
> step of registering as [party members] (which is still very minimal compared to most other
> democracies where party members must usually pay fees, attend meetings,
> etc). This is the advantage of not having a Bill of Rights in Australia:
> that citizens other than Justice O'Connor get to put forward, even enact,
> their convictions about what a fair electoral system means...
>
> --------------------------------------------------------------------------
> Dr Tom Round
> BA (Hons), LL.B (UQ), PhD (GU)
> Research Fellow, Key Centre for Ethics,
>          Law, Justice and Governance (KCELJAG)
> Room 1.10, Macrossan Building, Nathan Campus
>          Griffith University, Queensland (Australia) 4111
> Ph:     (061 or 07) 3875 3817
> Mbl:    0438 167 304
> Fax:    (061 or 07) 3875 6634
> E-mail: T.Round@mailbox.gu.edu.au
> Web: http://www.gu.edu.au/centre/kceljag/
> --------------------------------------------------------------------------