[EM] more questions about PR allocation formulas

[Olli Salmi]

At 01:53 -0400 13.10.2003, James Green-Armytage wrote:
>I thought that Newland-Britton was the
>method of doing fractional transfers of surpluses rather than transferring
>randomly selected ballots

Newland-Britton rules
http://www.cix.co.uk/~rosenstiel/stvrules/

>Also, I got another message
>that seemed to be saying that it is very uncommon for actual list systems
>to use largest remainder forms of allocation such as Hare and Droop, and
>it is far more common to use "highest average," divisor-based methods such
>as D'Hondt and Saint-Legue. Can anyone confirm or deny this?

This is probably the case. The Largest Remainder method is easy but 
it produces paradoxes. Germany uses Hare-Niemeyer because they find 
d'Hondt too biassed. Yet they have a legal threshold to take back the 
advantage. They seem to be moving to Sainte-Laguë because they don't 
like the paradoxes.

In the Sainte-Laguë countries in Scandinavia, d'Hondt is used within 
parliaments and councils for the election of committees. It occurred 
to me some time ago that the reason for this could be that if you 
know how the votes are going to be distributed, you might be able to 
exploit the rounding rule to your advantage.

Olli Salmi