[EM] 10/30/99 - Comments on Schulze's example.

[Markus Schulze]

Dear Donald,

you wrote (30 Oct 1999):
>     If your example is an One Area election, like the council of a small
> city, then I would suggest that the best thing the city could do is to
> increase the number of members to the even number of six. In that way the
> 50:50 results will be more proportional to 52:48. (52 is closer to 50
> than to 60).
>     The council would need a conclusive majority of four of six votes in
> order to pass any measure. The two parties would need to work together for
> the good of the city, which is the way it should be.
>     If your example is only one district of many, then the best these
> people can do is to change the system to Open Party List, with all the
> districts combined together in One Area.
>     If the people require a 'member link' then they can use a MMP type of
> election method, but I digress.

Of course, there are many ways to solve this problem.

One way would be to calculate a Condorcet ranking for every candidate and
to eliminate successively that not already elected candidate with the
lowest Condorcet ranking instead of that candidate with the lowest actual
number of first preferences. Another way would be to give bonus seats to
that party that has got a majority of first preferences but that hasn't
got a majority of seats; this is actually done in Malta. But -to my opinion
and to the opinion of many other people- the best way is to use the Droop
Quota instead of the Hare Quota.

You wrote (30 Oct 1999):
>     You have implied in your example that the 52% block of votes should
> receive three of five seats. If so, then suppose this 52% of the voters
> split into three parties before the next election. I am justified in asking
> you to take this step because in reality all the elections around the world
> that use the Droop have more than two parties, some have a lot more than
> two. Your example of only two is not realistic.
>     Do you believe these three parties should get three of five seats?
> Will not happen if we use Droop. The 48% party will get three seats.
>     Wow, isn't that something, Droop made it possible for a minor party to
> win a majority of the seats. David Catchpole was correct, Droop is God
> Almighty.
>     I content that this is the real reason for the use of the Droop Quota.
> It is for the purpose of helping the largest political party to gain a
> majority, even if that party may have less than a majority in popular
> votes. Droop is legal corruption.

It is not true that Droop favours larger parties.

If there were one conservative party and ten socialist parties and
if 48% of the voters prefer every conservative candidate to every socialist
candidate and 52% of the voters prefer every socialist candidate to every
conservative candidate, then the socialist parties would get three of the
five seats and the conservative party would get only two of the five seats
although the conservative party is very much larger than the other parties.

You wrote (30 Oct 1999):
>     Getting back to your example, what you have presented is a good
> example of a large group lowering their vote numbers on their candidates in
> order to scam extra representation for their party. This is possible to do,
> STV has this weakness, this flaw, it can be corrupted. We know this. It is
> another good reason not to use STV.
>     STV should be rejected, it is not proportional to the first choices,
> plus it has been corrupted too much.
>    If you are looking for proportionality of the first choices, then you
> need to look to Open Party List or MMP or some hybrid of STV/Open Party
> List as your multi-seat election method.

I am surprised about the sudden change of your opinion about STV. A few
days ago you were an enthusiastic supporter of STV and now you say that
STV has "flaws" and is "corrupted" and "not proportional." What has caused
this change of your opinion?

I don't agree with your pessimistic statements about STV.

Markus Schulze