STV-MD PR

[Marcus Ganley]

per Steve Eppley  Wed, 31 Jul 1996 18:51:29 -0700 (PDT)

>plain STV PR isn't good for electing a huge number of seats ... it's too many 
>candidates to rank, and it's not safe to rank only a few because some of 
>one's weight will probably get wasted.
I agree that the major problem with STV is that it is difficult to 
use when the electorate magnitude (ie the number of legislative seats on offer 
from that electorate) is high.  At the last double-dissolution election in Australia 
(an exceptional election inspired by a deadlock between the upper and lower 
houses in which the whole Senate, rather than just half, faces relection) each 
state had an electorate magnitude of 12 - this became close to unworkable.  Of 
course the problem was heightened by the fact that in Australian Senate 
election any truncation DOES invalidate your vote (see Commonwealth 
Electoral Act 1983).  Without this unnecessary element of compulsion the 
problem would not be so great.  

>I've a theoretical question:  Suppose a 20 seat body is being
>elected using STV from two districts each with an equal population.
>Should each district elect 10 seats?  
According to Taagepera & Shugart under any system of PR as the magnitude increases the proportionality increases.  However proportionality is disctorted where there are less than five seats being contested.  Thus while it is necessary to keep the elect
The other concern that arises with multi-distict STV is 'gerrymandering'.  That "gerrymandering" is possible with multi-member electorates is evidenced by the experience in Ireland prior to 1981 (see Mair 1986: 299-307).  Mair identifies three methods 
While I tend to support the use of STV with multiple districts each returning 5-9 candidates another option is to allow for intra-party preference with a party list system.  
While the essential characteristic of list systems is "fair shares for the parti
es" (Lakeman 1974: 92) they can allow for sophisticated expression of preferences between candidates (see Katz 1986).  In Luxembourg voters have as many votes as there are candidates to elect (ie as many votes as the district magnitude).  They may simp
Arklay, T. and C. Griffith. 1990. "Citizens for Democracy Submission to EARC". In Legislative Assembly Electoral Review - Stage 1 (Voting System).  Public Submissions, Vol. 1. ed. Electoral and Administrative Review Commission. Brisbane: Government Pri
Katz, Richard S. 1986. "Intraparty Preference Voting". In Electoral Laws and their Political Consequences, eds. Bernard Grofman and Arend Lijphart. New York: Agathon Press.
Lakeman, Enid. 1974. How Democracies Vote, A Study of Electoral Systems. 4th ed. London: Faber and Faber.
Mair, Peter. 1986. "Districting Choices under the Single-Transferable Vote". In Electoral Laws and their Political Consequences, eds. Bernard Grofman and Arend Lijphart. New York: Agathon Press.
Round, Tom. 1990. "Proportional Representation Society Submission to EARC". In Legislative Assembly Electoral Review - Stage 1 (Voting System).  Public Submissions, Vol. 3. ed. Electoral and Administrative Review Commission. Brisbane: Government Printe
Taagepera, Rein and Matthew Soberg Shugart. 1989. Seats and Votes. New Haven: Yale UP p 113 

>Or should each district's apportionment depend on the number of 
>voters who turn out?  (One simple way to do this would be to tally 
>all the ballots together, using a quota based on the total seats and total votes.)

This is an interesting suggestion.  I am not sure how it would work 
exactly.  I need to spend some more time pondering it.  Being an 
Australian I don't think much about turnout - again unneccessary 
compulsion in the Australian electoral system - vote or be fined!

The problem I can see is that you could find yourself with 
unproportionally small electorate magnitudes where turnout is low, 
and even numbers in other places.  This may not be so much of a 
problem if for example you had a legislature of 243 members and 
nominally assigned 9 seats to each of 27 electorates.  Then based on 
turnout you reallocated each of the 243 seats to the 27 electorates.  
It would be exceptional for an electorate to then have less than 5.  

Another problem would be ensuring that it is easy to register and 
turnout to vote.  You are probably tired of hearing foreigners harp 
about turnout in the States but I would be incredibly hesistant to 
introduce your proposal until it was clear that all barriers to 
registration and voting are lifted.  Among other things I have real concerns 
about the whole concept of voting on a working day - regardless of 
legislative protections some employers are still going to pressure 
potential voters into not voting.   I know these points are 
tangential but they are also important.  

Theoretically your proposal appears sound, in practice it could be 
problematic.  Incredibly interesting though.  Thanks for putting it 
up!!!