[EM] Conceiving a Democratic Electoral Process

[Kristofer Munsterhjelm]

On 07/16/2012 12:23 AM, Fred Gohlke wrote:
> Good Afternoon, Kristofer
>
> re: "Strictly speaking, clones are candidates that are so alike
> each other that every voter ranks them next to each other
> (but not necessarily in the same order)."
>
> and
>
> "More generally speaking, a clone could be considered a
> candidate that's very close to an already existing candidate
> and whose presence changes who wins."
>
> Thank you. That's a clear explanation.
>
> Even allowing for my general ignorance of the topic, cloning seems to be
> more significant for multi-party systems than for the two-party system
> that dominates U. S. politics.
>
> Nah. I guess that opinion is wrong. In the U. S., a third party probably
> can't avoid cloning some portion of a major party candidate. If so,
> eliminating clones probably increases the distance from a two-party
> system to a multi-party system. Anyway, wouldn't we be better served by
> conceiving a way to advocate the common interest instead of worrying
> about whether or not a clone will harm the parochial interests of
> partisans?

Well, in a working party system, Republicans (say) who don't like the 
direction the actual Republican party is taking could go form their own 
party. Since they're closer to R than to D, one might call them clones: 
everybody (or nearly so) who has a preference for Republicans first 
would put the "insurgent Republicans" above the Democrats; everybody (or 
nearly so) who has a preference for Democrats would put the Democrats first.

However, since Plurality isn't cloneproof, that splits the Republican 
vote. While the insurgent Republicans don't agree with the Republican 
party, they like the Democrats even less, and they don't want to give 
the election to the Democrats. So they usually don't form breakaway 
parties, and if they do, the voters won't vote for them for the same reason.

This (among other things) makes the party system worse than it could be, 
and might be why Plurality leads to two-party rule. As a contrast, IRV 
is cloneproof and doesn't start to behave oddly until the breakaway 
party becomes large enough. Instead of Plurality's two-party rule, 
Australia, which uses IRV, has two-and-a-half party rule.

Maybe Condorcet could lead to multipartyism even with single-winner 
districts - this would depend on whether it can deal with multiple 
viable parties without denying parties' growth into viable alternatives 
(as Plurality does) or starting to behave strangely (as IRV does). I 
think it could, but I have no direct proof.

In any event, Plurality's lack of clone resistance makes it really hard 
to start third parties - and so would make party-based government look 
worse than it could be. There's nothing inherent in the concept of 
parties that limits the voters to choosing between only two options. The 
system does that.