[EM] PR-STV with approval based elimination

[Jonathan Lundell]

On Apr 30, 2009, at 9:24 AM, Raph Frank wrote:

> A nice feature of PR-STV is that it still meets the Droop Criterion no
> matter who you pick to eliminate in rounds where no candidate meets
> the quota (assuming you don't eliminate an elected candidate).
> However, there doesn't seem to be much discussion on using a better
> elimination method.
>
> The eliminate worst candidate rule, means that PR-STV collapses to IRV
> in the single seat case.
>
> A simple modification would that voters cast 2 ballots, an approval
> ballot and also a ranked ballot.
>
> The method would proceed as normal except that in elimination rounds,
> the least approved non-elected candidate would be eliminated.
>
> The means that assuming that voters only approve candidates from
> within their own faction, then the intra faction election method would
> be approval voting, even though PR-STV is used to allocate ensure that
> the result is proportional.
>
> Another option would be to have the voter submit a ranked ballot and
> also have an "approval threshold" candidate.  All candidates ranked
> higher than that psuedo-candidate would be considered approved.
> However, that gives voters less freedom to approve across party
> boundaries (or less freedom to interfere with who is elected for other
> parties, depending on your viewpoint).
>
> It may have the effect of electing the most centerist of each  
> party's candidate.
>
> In the single seat case, this would yield a hybrid approval/IRV
> method.  If no candidate gets a majority, the least approved candidate
> is elected.  An approval winner who is also the condorcet winner is
> guaranteed to win.


The problem with these approaches (a problem, anyway) is that they  
abandon later-no-harm. That seems a rather high price to pay.