[EM] voter strategy & 2-party domination under IRV voting

[James Gilmour]

Juho Laatu  Sent: Friday, August 19, 2005 7:32 AM
> 
> I continued the chain of thoughts a bit.
> I must add one item in my list of good sides of two-party systems. That 
> is the fact that there will be a change, and the change will be considerable. ;-) 

I would add several important caveats here.  A two-party system is OK if it fairly represents the wishes of the
electors.  In Malta they now have a very strict two-party system and we know that's what the electors want because the
Maltese Parliament is elected by STV-PR.  When STV was first introduced (1921) more parties did get significant support,
but over the years the VOTERS have changed the system from multi-party to two-party.  The problem in the UK (and in many
other countries and states) is that a two-party system is the result of a defective voting system (usually FPTP in
single-member districts  -  as a British legacy!) that does not at all fairly reflect the wishes of the voters.

Sometimes there will be change (UK governments 1945 - 1979), but sometimes there not be any change (UK governments 1979
- 1997).  That depends on the sensitivity of the voting system.  FPTP is extremely insensitive to small and modest
changes in votes, but then grossly exaggerates any swing (the nonsense of David Butler's 'cube law').  The worst
manifestation of this is in the block-vote (multiple-X vote) in 3-member districts used in local authority elections in
England.

My own view is very simple: if a parliament or legislature or council is supposed to be representative, then the first
requirement of the voting system is that it produces a representative result.  If, on the other hand, the prime
requirement is single-party majority government, representing the largest minority if no party has majority support,
then it is not difficult to devise a voting system that will consistently produce such a result.  The electors need to
make up their minds about the kind of "democracy" they want.  I know opinion polls show they want both of these
incompatible outcomes, but the issues have to be argued through and decisions made about the objective.

> I'll be one step more concrete and describe one interesting line of 
> development for the multiparty systems. In order to allow and  make the 
> parties move, and hear and reflect the true wishes of the voters one 
> could support dividing the parties in smaller units with some small 
> ideological (or other, like regional) differences. Technically the idea 
> is to allow structured groupings in the elections. Instead of 
> having only party X in the elections the party could have subgroups 
> like X/right-wing, X/greenish, X/women, X/northern,  X/against-the-war, 
> X/join-the-EU. This way politicians would be more tied to following the 
> wishes expressed in the elections and it would be harder to continue 
> the same old policy after the pretty speeches before the elections :-). 

I believe that STV-PR could provide this without such formal subdivision, if it is what the voters want.  Of course,
parties want all their candidates to be "on message", but the reality is that all parties are coalitions and all parties
have their "right wings" and "left wings" or some similar diversities of views on some important issues.  Without being
cynical, I find some marketing jargon is helpful here: the parties need to explore the "segmentation" of the "vote
markets" in which they are vying for support.  If there is a "green vote", do we have the "green wing" among our team of
candidates?  etc, etc, as dictated by the local "vote market" in each multi-member STV district.  That can only make the
parties more responsive to the electors.

> There are voting systems that support alliances of parties but I'm 
> unaware of any systems that would allow this kind of hierarchical 
> structures.

I would not be in favour of formal hierarchical structures because all such formal structures tend to restrict voter
choice and voter representation.  I prefer to encourage the parties to offer the widest appropriate diversity of choice
to the electors and then leave the voters to choose the representation they want through the operation of a sensitive
voting system.

> My starting point is open list methods where voters vote individual 
> candidates (that belong to some grouping, that may belong to some 
> bigger grouping etc.).

The problem with all party list systems (including open lists), is that you cannot ensure proportionality WITHIN the
parties unless the votes are transferable between candidates within the parties.  If you are prepared to introduce that
complication, you may as well simplify the system and use the well-recognised STV-PR where the voters are completely
free to determine the patterns of transfers as they wish.

> Yes, STV style systems do not exclude parties although they allow also 
> the flexibility of living without strict political groupings. This 
> feature is a nice addition to the more traditional systems. On the 
> other hand political groupings have also their value in that voters do 
> not need to know the personal opinions of many candidates whom they 
> should list in their vote but they can rely in on the well known 
> ideological groupings.

My point was not that "STV style systems do not exclude parties", but rather that parties are not less apparent and no
less active where STV-PR is used for public elections.  In Malta, STV functions like a very strict party system, but
with the important difference that all the elected members are the candidates chosen for their personal support among
the voters.  In Ireland, there is a great deal of cross-party voting as the voters give a high priority to features
other than party membership when deciding who will best represent them.

Regards
James Gilmour