[EM] California Dreamin'

Dr.Ernie Prabhakar drernie at radicalcentrism.org
Wed Aug 25 06:56:53 PDT 2004

Dear Election Methods,

Despite Arnold's best efforts (which, frankly, are better than anyone 
else has done here for decades) California appears to be in the grip of 
a perpetual governance crisis.  The result is that at least one 
well-known columnist is calling for radical reforms, including 
proportional representation (PR).  So, I figured this is a good time to 
map out a structure for a legislature that reflects everything I've 
learned from this group.  If things go well, I hope to start shopping 
the idea around with other local reform groups (suggestions welcome!).

My assumption is that California is ready for PR and Condorcet, but not 
proxy democracy; perhaps arbitrary, but I have to start somewhere.  The 
goal, such as it is, would be to get something approved by 2010, and 
implemented by 2012.  I welcome your comments and suggestions.  I'm 
especially weak on PR allocation formulas (does anybody recommend 
anything besides STV?), so I'd particularly appreciate advice about 

Yours truly,
-- Ernie P.

Reengineering California: Towards A 21st Century Legislature
Draft 1, 8/25/2004 Ernest Prabhakar <DrErnie at RadicalCentrism.org>

My vision is to have a bicameral legislature, with an PR lower-house 
(Assembly) to write bills and a Condorcet single-winner upper-house 
(Senate) to edit them.   The idea is that the Assembly would draft 
bills, but that amendments proposed by various factions would be 
separate items rather than in-place changes.   So, for example Bill 42 
would have amendments A, B, C, and D attached.   This would take 
advantage of the energy, diversity and creativity of PR, and provide a 
low barrier-to-entry for factions to make their voices heard.

By contrast, the job of the Senate is to filter out the various ideas 
and find the optimal compromise.   Debate in the Senate would be 
focused on simply identifying the interesting set of options, e.g., i) 
42, ii) 42AB, iii) 42AD, iv) 42C. Again, it would be a low barrier 
(15%?) to nominate an option; they'd of course use Condorcet to vote 
amongst the options, so more choices wouldn't hurt.  In case of a 
cycle, the whole Smith set would go to the governor and he/she could 
pick one.

Crucially, though, the Senate would not be allowed to write new 
language.  This should eliminate a lot of the sleaziness (and time 
wasting) that currently goes on in conference committee, as well as 
make the Senate focused on legislative review (something California 
sorely lacks at the moment).  I'm toying with the idea of making the 
Senate term eight-years, with no re-election option, to diminish their 
need for campaign fundraising; the flip side is that I'd make them easy 
to recall, say by a white-ballot vote every two years.

For the Assembly, experts say we need around 300 representatives, which 
I figure we'd group into 15 macro districts of magnitude 20; thus, 
groups representing more than 5% of the electorate would have a shot at 
getting one candidate into the Assembly; again, somewhat arbitrary, but 
I don't want to completely eliminate locality, nor do I think really 
fringe groups need a voice.  Actually, most new groups have localized 
strength anyway, so this would in fact make it easier for them to get a 

At the lowest level, we'd have 300 micro-districts generated by a 
non-partisan panel, or an impartial algorithm based on 
compactness/minimum (weighted) circumference.   Five of these would 
make up a Senate district (hence 60 Senators), and four of those give 
us our macro districts (size 20).   Assembly voting would be handled 
via single transferable vote (STV). Ideally, each party would nominate 
20 people, all of whom live in different micro-districts; smaller 
parties could nominate less, but be required to delegate surplus votes 
(in advance) to other parties.   Voters would have the option of simply 
voting for party list in order of preference (necessary for large 
magnitude), though they could also specify (up to 20) individuals. 
Given the lack of a spoiler effect, there would be no state-sponsored 
primaries; parties could choose their own methods for nomination, 
making it easier for candidates to only run in the general election.

Seat allocation would be a bit tricky, given my desire for locality.   
I'm imagining general rules like:
- candidates with the most individual votes have preference to be 
elected for 'their' micro-district
- it is more important to preserve proportionality than to assign a 
candidate to a given district
- all else being equal, smaller parties should be given their locality 
assignments first; meaning large-party candidates in their district 
would be the first dropped from the party list

The four Senate districts within a macro district would rotate 
elections every two years (assuming eight-year terms).   As a bonus, 
the Four Color Theorem ensures that no two adjacent districts need to 
be elected at the same time (except in case of a recall).

As part of this effort, there would also need to be electoral reform 
for statewide offices (based on Condorcet), as well as lobbying reform 
(based on the idea of real-time transparency), but that is outside the 
scope of this paper.


More information about the Election-Methods mailing list