[EM] Even simplier anti-gerrymandering rule

[mrouse1 at mrouse.com]

> At 7:02 AM -0800 11/6/06, <mrouse1 at mrouse.com> wrote:
>>A simple solution to gerrymandering is to have the voters pick which
>> districting method they want. Every recognized political party in a
>> state could offer their own version (subject to certain restrictions,
>> like simple closed curves and equal populations in each district), and
>> voters could pick which one they prefer in a general election. That
>> districting method would then be used for the *following* election.
>
> Sounds more like a gerrymandering guarantee, at least if there's a
> majority part at the time of the districting election.
>
> How about this: same rule, but a voters are prohibited from voting  for
> the rule advanced by their own party (ignoring the problem of
> opportunistic party-switching). A variation on the old
> candy-bar-sharing algorithm.


That method would work, but I think you are underestimating the impact of
centrist/undecided/nonpartisan voters. Some people vote straight
Republican or Democrat tickets, but when you have everyone in the whole
ungerrymandered state deciding on which districting map to use, the result
is likely to be less partisan than the current method.

Plus, you could introduce ranking methods into the election this way --
rank the maps in order of preference, and then pick the winner via some
Condorcet method. Minor parties with a good map could be a good compromise
second choice for a majority of voters, and it would give them exposure
(If you like our maps, you'll love our candidate! or some such
advertisement).

Michael Rouse
mrouse1 at mrouse.com