[EM] California Dreamin', Take 2

James Green-Armytage jarmyta at antioch-college.edu
Wed Aug 25 20:17:59 PDT 2004

Dear Ernie,

	My personal opinion is that bicameralism is unnecessarily bulky for a
state legislature. So rather than re-conceptualizing the relationship
between the two chambers, I'd rather just condense it into a single
chamber. I think that people have enough trouble keeping track of what is
happening in state government as it is, and I think that having two
separate chambers makes this problem twice as bad. 
	I would like to elect the single chamber via STV-PR. For district
magnitude, 7 or 9 on average sounds fine to me. 
	I don't think that it is too attractive to elect a legislative chamber
via Condorcet's method, even if they are only deciding on legislation
rather than writing it, because single-winner districts can lead to
disproportionality, and encourage gerrymandering. For example... Say there
are 100 seats in the chamber. The Democrats have a 100%-0% majority in 49
of the districts, and the Republicans have a 51%-49% majority in the
remaining 51 districts. The Democrats have 73.99% of the vote, the
Republicans have 26.01% of the vote, but the Republicans have majority
control over the chamber. It's an extreme example, but you get the general
idea. Single-winner-district legislatures are usually bad news, I think.
	I know that you are very concerned with centrism, but I do think that a
good STV-PR assembly would still hinge on the center. Some of the
representatives would probably be rather distant from the center, but in
order to pass legislation one would need to build a majority coalition
which includes the center.
	If you still feel the need to inject more centrism into the system, I
would probably just suggest a "parallel" system within one chamber (as in
Russia, Japan, etc.) rather than creating separate chambers. That is,
elect most of the seats (80%?) via STV-PR, and the remainder by Condorcet.
Thus, you have a mostly-proportional assembly that is somewhat weighted
towards the center. I'm not personally convinced that this would be
preferable to just a straight STV-PR legislature, but I would prefer it to
a bicameral system. I think that Rob Loring is an advocate of STV-PR /
Condorcet parallel assemblies.
>Despite Arnold's best efforts (which, frankly, are better than anyone 
>else has done here for decades) 

	Feel free to elaborate on these (onlist or offlist, as you prefer).

>The optimal district magnitude 
>for PR is usually considered around seven (7), 

	Considered by whom?

>giving a 90% confidence 
>that a voter would get a candidate they agree with.  

	90% chance that they will get their first choice, or what? Wouldn't that
just depend on the ratio between candidates and seats?

>In case of a 
>cycle, the whole Smith set would go to the governor and he/she could 
>pick one.

	Why not use the iterative pairwise procedure which I proposed for small
group voting, at
	Also, in a small group, the Smith set (or minimal dominant set) may
differ from the Schwartz set (union of minimal undominated sets). For an
explanation of the difference between these sets, see

my best,

More information about the Election-Methods mailing list