2011/7/4 James Gilmour <[email protected]> > Kathy Dopp > Sent: Monday, July 04, 2011 2:53 PM > > > However, either the election method used within each party to > > > determine the list orders would be majoritarian (in which case the > > > system isn't proportional beyond the party level), > > > > Plurality is how it is done I believe. To have PR within the > > party would require some sort of party primary system I > > suppose to determine which candidates are on each list in the > > general election for each party. > > This suggestion misses the point. For any voting system to give full > effect to proportional representation of the voters, the > selection of the candidates to take the seats won by a party must be > decided by those who vote in the actual public election - not > decided by any kind of party primary. After all, the party primary (before > the public election) has already decided who should be > on the party's list and has ordered that list. > > > The nice feature of existing party list methods is that it > > allows the election of a large number of candidates to a > > large national body of legislators without requiring voters > > to rank individually a huge number of candidates. This makes > > the job for voters and election administrators much easier > > than asking voters to rank from among a huge number of > > candidates. > > But it is precisely this "nice" feature of most open-list party-list > systems that causes the failure of such systems to produce > proportionality WITHIN parties. > > As I said in my last message, asset-like systems can let you have your cake and eat it, if you trust your favorite candidate to agree with you in ranking other candidates. This is fundamentally different from trusting your party, because your "favorite candidate" in asset-like systems could, in principle, be arbitrarily close to you - even BE you, if you're willing to give up vote anonymity, and if the system allows this extreme. Most systems will put some limits on this, but still, they are far closer to this extreme than any party list system. Also, there is no need to stay within the arbitrary bounds of any party; a candidate can have affinities based on ideology, so candidates at the fringes of their party (including the centrist fringes) have full freedom.
> If you are going to do this properly, to produce a within-party PR result, > the voters for each party would have to mark preferences > against the candidates in their chosen party's list (not necessarily all > candidates, depending on the system you choose). And then > you would need to use STV-PR (or something like it as you don't like STV) > to determine which candidates should take the seats > allocated to each party. No such system could be precinct-summable, but > that is not a priority for everyone. > I disagree about the "no such system" statement. I myself have worked out an unpublished system which is not perfectly droop-PR, but is a ~99% approximation thereof; and which is complicated, but still 2n² summable. It's not worth sharing the details here, but, having gone through the exercise, I believe that it should be possible to do better than I did. > > And as has already been said, if you are prepared to go the bother of > counting what is in effect a separate PR election WITHIN each > party, why not go all the way and apply your chosen PR system to all > candidates across all parties? That would give the voters real > choice and would also avoid completely the problem of entrenching the > political power of the parties' machines. > Agreed. > > James Gilmour > > > ---- > Election-Methods mailing list - see http://electorama.com/em for list info >
---- Election-Methods mailing list - see http://electorama.com/em for list info
