<html><head><style type="text/css"><!-- DIV {margin:0px;} --></style></head><body><div style="font-family:times new roman, new york, times, serif;font-size:12pt"><DIV>Marcus,</DIV>
<DIV> </DIV>
<DIV>Kristopher wrote:<BR>"That MMP method would have to use some kind of reweighting<BR>or those voters who got their way with regards to the constituency <BR>members,..."<BR></DIV>
<DIV>And you responded:</DIV>
<DIV>"Wow, that's exactly what I have proposed recently for<BR>an STV-MMP system in Berlin."</DIV>
<DIV> </DIV>
<DIV>I don't completely understand your papers, but your suggested<BR>system seems to consider the individual voter's district vote and</DIV>
<DIV>party vote separately (i.e. each without regard to the other) so<BR>I can't see how you "reweight voters" or have a completely fair<BR>and satisfactory "top-up" system.<BR></DIV>
<DIV>Wouldn't it still be the case that the district voter with no or only <BR>very weak preference between a candidate nominated by a party</DIV>
<DIV>that is sure to qualify and an independent does better to vote for </DIV>
<DIV>the independent?</DIV>
<DIV><BR>I would prefer some quasi-STV system that uses a jurisdiction-wide<BR>uniform quota, something like this:</DIV>
<DIV> </DIV>
<DIV>Voters (strictly?) rank from the top candidates standing in their</DIV>
<DIV>district, and on the same paper also strictly rank from the top</DIV>
<DIV>lists of candidates for the jurisdiction-wide (quasi )top-up and</DIV>
<DIV>also may rank candidates within lists they rank. In both cases</DIV>
<DIV>truncation is allowed.</DIV>
<DIV> </DIV>
<DIV>Select a uniform quota and use to elect candidates in the districts</DIV>
<DIV>STV-style.</DIV>
<DIV> </DIV>
<DIV>Downweight each voters' top-up vote according to how much their</DIV>
<DIV>district votes contributed to winning candidates' quotas.</DIV>
<DIV> </DIV>
<DIV>Use these reweighted votes to elect list candidates thus:<BR>Eliminate all lists that don't qualify and IRV-style distribute preferences.</DIV>
<DIV>If the remaining list with the least votes doesn't have a quota, likewise</DIV>
<DIV>eliminate that list and so on until all remaining lists have at least one<BR>quota. Each list then wins as many seats as it has quotas.<BR></DIV>
<DIV>Then with the ballots that supplied each list's quota/s, use (regular) </DIV>
<DIV>STV to determine which individual candidates fill the list's seats.<BR><BR>(By "qualify" I mean qualify according to the German constitution.)<BR></DIV>
<DIV>A more sophisticated version might include some way of handling lists'</DIV>
<DIV>surplus votes.</DIV>
<DIV> </DIV>
<DIV>Candidates should not be allowed to stand in more than one district,</DIV>
<DIV>but should be allowed to stand in both a district and on a list.</DIV>
<DIV> </DIV>
<DIV>Without being restricted by the German constitution the "top-up"</DIV>
<DIV>part could use the same method as the districts. If the jurisdiction as a </DIV>
<DIV>whole could qualify as a "district" then perhaps this version wouldn't <BR>transgress the German constitution.<BR></DIV>
<DIV>For STV systems in general I don't see any great positive point in having</DIV>
<DIV>the numbers of seats for each district fixed before the votes are cast.<BR>Why not wait until after and make the apportionment between districts</DIV>
<DIV>based on the numbers of votes cast? That would tend to make the overall</DIV>
<DIV>result more proportional, and districts with higher turnouts would be rewarded\</DIV>
<DIV>with more seats.</DIV>
<DIV><BR>Chris Benham</DIV>
<DIV><BR> </DIV>
<DIV><BR> </DIV>
<DIV>Marcus Schulze wrote (<EM>Sun Jul 6, 2008):<BR></EM></DIV>
<DIV>Dear Kristofer,<BR><BR>you wrote (6 July 2008):<BR><BR> > I've been reading about the "decoy list" problem in mixed<BR> > member proportionality. The strategy exists because<BR> > the method can't do anything when a party doesn't<BR> > have any list votes to compensate for constituency<BR> > disproportionality. Thus, "cloning" (or should it be<BR> > called splitting?) a party into two parties, one for the<BR> > constituency candidates, and one for the list, pays off.<BR> > But is it possible to make a sort of MMP where that<BR> > strategy doesn't work?<BR> ><BR> > That MMP method would have to use some kind of reweighting<BR> > for those voters who got their way with regards to the<BR> > constituency members, I think, because if the method<BR> > just tries to find correlated parties, the party could<BR> > theoretically execute the strategy by running all the<BR> > constituency candidates as
independents. What kind of<BR> > reweighting would that be? One idea would be to have a<BR> > rule that says "those with say x about the constituency<BR> > vote gets 1-x in the list vote". Then vary x until the<BR> > point of party proportionality is found. No matter what<BR> > party someone who makes a difference with regards to the<BR> > constituency candidate chooses, his vote loses power<BR> > proportionally, and thus decoy lists wouldn't work.<BR><BR>Wow, that's exactly what I have proposed recently for<BR>an STV-MMP system in Berlin. Please read these papers:<BR><BR><A href="http://m-schulze.webhop.net/schulze4.pdf"><FONT color=#810081>http://m-schulze.webhop.net/schulze4.pdf</FONT></A><BR><A href="http://m-schulze.webhop.net/schulze5.pdf"><FONT color=#810081>http://m-schulze.webhop.net/schulze5.pdf</FONT></A><BR><BR>Read especially page 3 of paper "schulze5.pdf".<BR><BR>Markus Schulze<BR></DIV></div><br>
<hr size=1>
Find a better answer, faster with the new Yahoo!7 Search - <a href="http://au.rd.yahoo.com/mail/taglines/au/y7mail/default/search/*http://au.search.yahoo.com/search?p=pete+murray&fr=au_ssw&p1=search&p2=other&p3=au&p4=tagline" target=_blank>Start Here</a>.</body></html>