[EM] Re: bicameral design poll

James Gilmour jgilmour at globalnet.co.uk
Mon Jan 5 19:14:39 PST 2004

> > > Interesting question.  I agree with your answer except for
> > > the size of the districts.  According to Duverger, a district 
> > > size of 10 supports 11 parties. 
> > 
> > What does this statement mean?  I am not aware that there is any 
> > direct relationship between district magnitude and the number of 
> > active or contesting parties.
> Duverger proposed this theory and it is well known in the 
> political science community.  You should be able to find more 
> details in academic articles and textbooks.

Thanks for the pointer.  I found Duverger's own statements like "the simple-majority single-ballot
system favours the two-party system".  He was much more cautious in his exposition than some of
those who came after him.  In the publicly available webpages I found nothing to suggest that
districts of N seats will "support" N+1 parties.  And there are plenty of examples to show that
factors other than district magnitude can be the main determinants.  Italy had many politic parties
when it used majoritarian voting; the numbers of parties did not increase when party list PR was
introduced.  With STV-PR Malta (5-member districts) ahs moved from a multi-party system to a
two-party system.  In contrast, in Northern Ireland with STV-PR (6-member districts) a multi-party
system have been maintained for more than 30 years.

> > 4 or 5 may be the ideal size in a particular situation, but 
> "reducing 
> > the number of candidates" is not a good reason for making 
> such a decision.
> IMHO, the number of candidates is an important consideration 
> in designing an electoral system.

It may be important, but if you take care of the more important features in designing the electoral
system, the number of candidates will take care of itself.

> I think too many 
> candidates can overwhelm the voters and hinder meaningful 
> political dialogue.

True, but it is not a problem in practice if the other features of the voting system are right..
Scotland used STV-PR with district magnitudes varying from 3 to 10 without apparent problems to
elect its Education Authorities in the 1920s.

> 150 candidates for the election of the 
> governor of CA is clearly too much.

Unusual events like this prove nothing of general application.


More information about the Election-Methods mailing list