[EM] Why I think IRV isn't a serious alternative

[Kristofer Munsterhjelm]

Juho Laatu wrote:
> --- On Wed, 26/11/08, Kristofer Munsterhjelm <km-elmet at broadpark.no> wrote:

>> I suppose what I want is a combination of this. The voter
>> should have what he wants (which is to say, the method
>> should make it more likely that the median voter can get
>> what he wants), but if there is no selection among
>> alternatives, that can't happen. In other words, if
>> there is a two-party state and the parties don't care
>> about your issues, you're out of luck; and if they say
>> they care about your issues, only to turn around once
>> elected, you have little chance to do anything. A duopoly is
>> bad whether it's a political or economic duopoly.
>>
>> So I want the people to be able to get what they want, but
>> also the method to support the circumstances that ensure
>> that will be true in the future as well.
> 
> I think this is one key to how also current multi-party systems could
> be improved. Many people "hate" the party structure since often the
> parties seem to be quite stagnated and deaf to the voices of reason. If
> this happens in a situation where the country is in a stable state
> without any risk of too rapid movements in the political structure, then
> it may make sense to encourage the political parties/structure to react
> better to the needs / development interests of the citizens.
> 
> Example 1. STV is a method that allows voters to influence very
> freely  on which candidates will be elected when compared to more
> party oriented methods. There may be some drawbacks in complexity
> (especially if there are many candidates) and lack of structure
> (candidates are not bound to programs).
> 
> Example 2. Subgroups withing the parties allow voters to influence
> more on which candidates will be elected when compared to basic party
> oriented methods. Expressiveness is more limited than in example 1.
> Groupings are more clear than in example 1.
> 
> In both approaches the main expected benefit thus is that these
> methods are supposed to make it possible to the voters to have a say on
> what direction the political parties will grow (example 2 focuses more
> on this), or allow elected representatives to form freely any kind of
> coalitions (=not necessarily bound to the official and limiting party
> policy) when making decisions (example 1 is more radical here).

I think that the second approach (that is, your first example) is the 
one to focus on, but that's probably because I think that parties 
shouldn't be made official parts of the system. If the parties find that 
independents are taking the seats, or that members lower on their lists 
get consistently ranked higher, then the parties should do something 
about it.

That could fail if parties have the general power to set the agenda, so 
that they can use their power to concentrate their power further, in 
which case one might just have to concede and let the voters have direct 
power over party matters instead.

As for the second approach, I'll say that STV is a good method (with 
sufficient seats), Schulze's STV-MMP is somewhat better (if more 
complex), but also that it's possible to make better party-neutral PR 
methods than STV. If my simulations are of any value, QPQ is one such 
method, but it's that by coincidence more than anything, I think, so 
there's still room for trying to find ways to design a better 
party-neutral PR method "from first principles", and for discovering how 
to do so.