[EM] the 'who' and the 'what'

[Kristofer Munsterhjelm]

Raph Frank wrote:
> On 9/26/08, Kristofer Munsterhjelm <km-elmet at broadpark.no> wrote:
>>  It seems this system would be more stable than I originally thought. Third
>> parties could run as parts of the Condorcet party without running much of a
>> risk, since they would otherwise get no votes at all. The defection danger
>> surfaces when the third parties have become sufficiently large from using
>> that parallel electoral system. Then a party that would win a plurality vote
>> but who isn't a Condorcet winner has an incentive to defect.
> 
> If the condorcet party winner can realistically claim to be one of the
> top-2, then it doesn't matter as he will defeat any challeger.  Both
> the 2 main parties would have to defect.
> 
> The question is at what level of support does this becomes self-reinforcing.

Not all parties may have chosen to join the Condorcet party in the first 
place. If so, multiple parties could draw away votes from the Condorcet 
party, even if only a single party defects. I think you're right about 
the more particular case, though, and that shows that a Condorcet system 
grafted on top of Plurality, where all the parties are members of the 
Condorcet "party", is very stable indeed. There's a problem if multiple 
parties defect (and also a problem if there is no CW).

Though if all parties are members of the Condorcet party, then probably 
they could agree to get rid of Plurality; or they could replace it with 
a runoff.

>>  Following that kind of reasoning, it would appear that conventional parties
>> have very little to lose by running Condorcet primaries instead of Plurality
>> primaries, more so if there's an open primary. (So why don't they?)
> 
> The current parties don't want to elect a condorcet winner, they want
> to elect a winner that is biased towards them.
> 
> The 2 candidates in a 2 party system have to balance support of their
> party with defeating the other candidate.
> 
> In the single issue case with voters ranging from 0 to 100, the 2
> parties pick at 25 and 75, but the condorcet winner is at 50.

I know that. But consider a party that uses Plurality internally for its 
primary. The party's center is 25, and the candidate closest to that 
center "should" be elected for the primary. However, we know that 
Plurality is flawed. Therefore, it may elect a candidate closer to 50 
(which is bad from the ideological point of view of the party, since it 
waters out the position of the candidate), or one that's closer to 0 
(which is also bad, because it makes the candidate less acceptable to 
the people).

So let's say there are two parties, left and right, that have centers at 
25 and 75 respectively. Left has decided to switch to Condorcet. Because 
of this, it elects the candidate at 25. Right doesn't care about 
Condorcet, and due to bad luck, candidates within the party split the 
vote so that the candidate that wins is closer to 80. Then Left wins 
because 25 is closer to 50 than 80 is. Therefore, it would seem that 
switching to Condorcet would benefit the party in that it would much 
less often get its victory stolen by the artifacts of the system.

That's what I mean: switching to Condorcet primaries should only improve 
the accuracy of the internal vote, with little disadvantage, even if 
such an accurate result for a party is more extreme than the accurate 
result for the people in general. So why haven't the parties done so?

Perhaps it's one of Michael Allan's answers: they don't know about the 
technology, or Condorcet has the potential to weaken the party elites 
more than they like, or internally split the party itself. I am not sure.

Note that I'm not talking about the major parties uniting under the 
Condorcet party, but the parties replacing their internal Plurality 
primarities with Condorcet primaries.

> Anyway, I would agree that an open primary would be key for the
> condorcet party.  In states with a closed primary can a party allow
> non-party members to vote if it wishes?  Would this block those voters
> from voting in their 'real' party?

I think that the voters have to be registered with the party in 
question. I don't live in the US, though, so I could be wrong. (Over 
here, we have party list PR and so don't need any registration)

> Another problem is actually getting the main candidates to
> participate.  I assume it would be legal to add them to the ballot
> without their permission?

Even if not, that's where the stability of the metasystem comes into 
play. A third party with little support has next to nothing to lose by 
joining the Condorcet metaparty (which would be a party just enough to 
let it run as one when the internal Condorcet election among the 
candidates from the real parties that are part of it has been decided). 
The worst thing that can happen is that the artifacts of Plurality will 
swing the other way so that they lose whereas they'd won otherwise; but 
if so, the party in question isn't a minor party by any account. Thus, 
you'd just have to wait for the alternate parties to all join the 
Condorcet party to amplify their individual strength; you don't have to 
force the parties to be part of the process.

Though now that I think about it, perhaps there's a problem with a 
Condorcet party. Say that parties A and B have decided to not run their 
own candidates, instead going through the Condorcet party. A wins the 
internal Condorcet election. Now B-voters could turn to the lesser evil 
instead of abstaining. This could result in unpredictable effects.

> Finally, turnout at the condorcet primary matters.  If only a small
> number of people vote, then it is much less evidence that the winner
> is the real condorcet winner.  One option would be to re-weight votes
> so that the result is representative.
> 
> If the consequences of the result of the vote is not massive, then
> there is little point in bothering to vote.  So, there needs to be
> some kind of boot-strap.

Indeed it does matter. The voters that support parties that have joined 
the Condorcet party would have to vote twice: once in the "interparty 
primary" (Condorcet election), and once in the general election 
(Condorcet party vs other parties). This may be cumbersome, but then 
again, the feeling of empowerement might offset it sufficiently that the 
voters aren't discouraged.

> Once the condorcet winner can credibilly claim to be one of the top-2,
> then the condorcet primary almost becomes the final election.
> Certainly, winning the condorcet primary would be a major boost to any
> candidate.

Once the CW can claim to be one of the top two, then the general 
Plurality election reduces to a majority election (party A vs the 
Condorcet party winner). If the CW is the true CW, he'll win outright, 
not just be boosted.