[EM] PR for ethnically polarized electorates

[Forest Simmons]

The "Ultimate Lottery" is one such method, but it satisfies only a weak
form of mono raise: improving the position of a party on some ballots will
not reduce that party's calculated proportional share to zero.

The conditional approval approach satisfies strict monotonicity: improving
the position of a party on some ballots cannot decrease the party's
calculated proportional share.

In the context we are considering, namely polarized electorates like those
of Iraq, Rwanda, etc. the main faction sizes are well known.  This fact
should enable moderate parties to publish recommended conditional approval
ballots, or allow voters to vote "below the line"


On Tue, Jul 1, 2014 at 2:00 PM, Kristofer Munsterhjelm <km_elmet at t-online.de
> wrote:

> On 07/01/2014 12:19 AM, Forest Simmons wrote:
>
>> If the voters don't want to strategize, then they just submit their
>> range ballots (for the parties) directly.  Let R(i) represent range
>> ballot R's rating of party i, and let S(i) be the social rating of party
>> i, in other words the average value of R(i) over all R.
>>
>> Then the single vote of ballot R goes to the party i that maximizes the
>> product R(i)*S(i).
>>
>> If the allowable range values are just zero and one, then this method
>> reduces to Martin Harper's vote assignment scheme.
>>
>
> Won't that be susceptible to one-way strategy?
>
> What I meant was, is there a method that can do what the strategic method
> does (given the negotiation data) but without any explicit negotiation data
> available? Something like DSV, although in this case, I suppose that the
> ballot with negotiation data (how much agreement you need to switch) is
> technically DSV.
>
> Or perhaps there is something that is similarly strategy resistant in this
> situation as MJ is in ordinary elections wrt Range.
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20140702/c5101d86/attachment.htm>