[EM] The responsiveness of Condorcet

[Dgamble997 at aol.com]

James Green-Armytage make two postings in support of Condorcet. The subject 
of the second posting ( centre dominance ) is the principle reason I consider 
it a flawed method. I realise that I've probably made every criticism that can 
be made of it in the course of the last month but my principle objection to it 
is that it has great potential to produce some very bad results in practice 
in certain real life situations.

When I first read about Condorcet one of the first things that I did was 
apply it to the party system of my own country ( Britain). I then fairly swiftly 
rejected it as a method because it seemed extremely probable that if applied in 
single member seats  it would give very bad results.

My first criterion for good results for an electoral system is that it 
allocates seats to parties in proportion to their vote.

In England there are three main political parties, two large ones - the 
Conservatives on the right and Labour on the left and a smaller third party - the 
Liberal Democrats   generally perceived by voters as being in the centre. The 
Liberal Democrats are by no means a turkey party just a smaller party perceived 
by voters as being in the centre.

In Scotland and Wales their are additionally the Scottish National Party- a 
broadly left wing party that wants an independent Scotland and Plaid Cymru ( 
Welsh for the party of Wales) which wants more power devolved to Wales and 
strongly supports the Welsh language and culture.

Additionally in proportional elections for the European parliament,  
Scottish, Welsh and London assemblies the Greens, Scottish Socialist party and UK 
Independence party have all obtained seats.

Ulster has a party system of its own with 2 Unionist parties, 2 Republican 
parties and the cross-community Alliance party.

In brief we have a much more complex system than the Republican/Democrat 
duopoly.

In many House of Commons seats in England the result looks something like 
this

Conservative 41%
Liberal Democrat 14 %
Labour 43 %
Others 2%

Assuming that all Conservative and Labour voters vote Liberal Democrat as 
their second choice ( in practice this wouldn't occur 100 %  but it is likely 
that to a great degree that is what would happen) the Liberal Democrat would win 
the seat. This result would probably be repeated in  many seats giving a 
grossly disproportional  outcome.

In an earlier posting I applied Condorcet to the 1997 General Election result 
for the 17 seats in the county of Kent. I made the assumption that 100% of 
Conservatives voters and 100% of Labour voters give a second preference to the 
Liberal Democrats and that the Liberal Democrats divide their support 50:50 
Conservative: Labour.

The voters were cast as follows:

Conservative 40.5%
Labour          37.2%
Liberal Democrat 17.0%
Others  5.4 %

Under First Past The Post (plurality) the seats were distributed 9 
Conservative, 8 Labour, 0 Liberal Democrat.

Under IRV the seats were distributed 7 Conservative, 8 Labour, 2 Liberal 
Democrat.

Under Condorcet the seats were distributed 1 Conservative, 3 Labour, 13 
Liberal Democrat.

I simply cannot support a system that gives a party with 17 % of the vote 77 
% of the seats. Whatever theoretical and conceptual benefits it may have I 
simply cannot support it.

I see the attractiveness of Condorcet ( and Approval Voting) as applied to 
America. They are the single seat methods which would make it easiest for a 
third party ( provided it could attract the second preferences of Democrats and 
Republicans) to win seats and break the duopoly ( in practice an extremely 
difficult task).

Condorcet ( or Approval Voting) would appear to be the best methods for 
changing the current American party system. However if you change the current 
system, it will by definition, become something else. The something else that 
Condorcet seems to favour is a Centrist monopoly.

David Gamble




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20030712/722ec591/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: KENTVOTES.xlr
Type: application/octet-stream
Size: 14336 bytes
Desc: not available
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20030712/722ec591/attachment-0001.obj>