[EM] Amnon Rubinstein's Proposal for Electoral Reform in Israel

Antonio Oneala watermark0n at yahoo.com
Mon Apr 24 16:07:30 PDT 2006

James Gilmour <jgilmour at globalnet.co.uk> wrote: > Doreen Dotan Sent: Sunday, April 23, 2006 2:42 PM
> 1) To separate the party lists we presently vote for into 
> regional slates and a general list thus: 
> Divide the country into 40 election districts with 2 
> representatives for each district. Representatives *must* be 
> residents of the district they represent. 
> The remaining 40 representatives are elected according to the 
> general count, from a separate list. A regional candidate may 
> not be on the national list. 
> This would give 120 Knesset members, only 1/3 of whom are 
> elected according to the present system, with district 
> representatives accountable to their electorate. 

I would NOT recommend any voting system that elected the members in two such different ways. It is a recipe for
potential trouble, particularly if the district representatives are predominantly from the two largest parties
(inevitable with 2-member districts) and the national list members are from the smaller parties, and especially so if
the larger parties from the government and the smaller parties are all in the opposition.  That's what we've got in
Scotland and that's why we have so much trouble with MMP in our Parliament which doesn't arise elsewhere (or at least,
not to the same extent).

All members of the Knesset (and the Scottish Parliament!) should be elected on the same basis and all should be directed
accountable to the local voters.

> It would also demolish the small, specific interest-based 
> parties that have often held entire governments up for ransom 
> with the threat of leaving a narrow coalition. 

Some adjustment in this direction in Israel would seem desirable (!!), but this proposal, dominated by 2-member
districts, restricts direct local representation to a wholly unreasonable extent.  With STV-PR you could have districts
electing, say, 5, 6, 7, 8 or 9 members, with size being related to the various "natural communities" and to population
distribution (urban v. rural).

James Gilmour

election-methods mailing list - see http://electorama.com/em for list info

I've never really been a fan of STV-PR.  It's still a system that's not independent of irrelevant alternatives, so in most states it squeezes out third parties and moderates, unless you expand the districts to an enormous size (in which case barely anybody bothers to rank candidates, and if you allow party line votes, they will probably simply check one of those, so it boils down to an innacurate party-list method).  

I'd prefer Simmons PAV method, but that has practically no recongnition worldwide, despite it's obvious merits.  For one, unlike STV affecting the amount of candidates that your party runs won't mean you win more or less seats.  For another, moderates won't be squeezed out by the same effect that causes them to be squeezed out in IRV.  Unlike STV, PAV has the advantage that you never have to betray your favorite in order to get another candidate into the first few rounds.  Besides all of that, it is obviously easier to rank.

The only real problem I see with it is that, while it is not a technically complex method, the amount of data involved is massive.  A single election in a 4-member constituency, under ideal conditions, could have the amount of election data that over 1000 normal approval elections have.  However, it would take a computer all of seconds after a voter had voted to compute all of this together, and then it could all be streamed to a central computer to pick the winner.  All that really matters is the actual votes the voter cast, so a hand recount could be performed with each counter having the aid of a computer.

So, here would be my proposal.  I'm a fan of bicameral legislatures, and I'd want one to be majoritrian, for stability, and one to be proportional, in order to allow minority voice in.  You could then divide the country into thirty-five constituncies, and have each one elect four persons through PAV every three years, and have them elect two at-large electors by normal approval voting, but give each at-large legislature six year terms, staggering them so that one is elected every three years.  That gives the legislative body with at-large members 70 members, and the body with proportional member 140.  The ballot should be completely non-partisan (I think partisanship is bad for the legislative process), and the legislative rules should be constructed to remove most possibilities for party discipline.  A body to form the boundraries of the districts should be formed, and it's purpose should be to make the elections in each constituency as competitive as possible.

I think the executive branch should be seperate, but electing straight may be a bad idea.  Possibly, you could split the legislative branch into a set of 10 or so "executives", half of them elected through PAV, the other half through districts.  Give them various some of the executive powers, and delegate the rest to an elected PM and cabinet, which may be composed of the members of the executive body.  This has never been tried before, I'm not sure of it.

Love cheap thrills? Enjoy PC-to-Phone  calls to 30+ countries for just 2¢/min with Yahoo! Messenger with Voice.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20060424/c9267504/attachment-0003.htm>

More information about the Election-Methods mailing list