Re: [EL] Redistricting answer

I call this the Big Sort Lie. Increased residential segregation -- if even
true, the Bishop's Big Sort book is based on county data and treats Los
Angeles the same as North Dakota counties with 100 people; I have not seen
recent work by Bruce Oppenheimer on this subject -- only makes a significant
difference in a few places where there are edge effects (i.e., San Francisco
or NYC). Those who have drawn districts know that congressional districts
are sufficiently large to bridge different communities and that residential
segregation aids in reaching target partisanship numbers. I agree with Nate
when district competition is NOT a criterion, I disagree with Nate when it
IS a criterion and it is not subsumed under other criteria (though Nate made
no claim on the second point, only the first, so I not really disagreeing
with him), and offer the 2009 Ohio redistricting competition as evidence.
There are some extremely non-compact districts out there that are not voting
rights districts -- such as IL-17, which was designed to be a safe
Democratic district -- so I do not believe that residential segregation
patterns are necessarily limiting factors on the number of competitive
congressional districts one might draw. However, I practically think
compactness should be among the criteria to balance against others like
district competition. I highly recommend getting some GIS software are
drawing some districts, as this will be very illuminating about what is and
is not possible.

There has been a good deal written on voting patterns of new vs. old
constituents with regards to the incumbency advantage, so I am not sure
there is much being missed on this topic. I will reiterate, there is a
difference between competitive districts and competitive elections. Antoine
is (unintentionally, I hope) blurring the distinction. I'm willing to give
up here and give competitive districts a new name, like partisan-balanced
districts. I agree that the optimal partisan gerrymandering strategy
includes shaking up the map -- particularly to force the other party's
incumbents to run against one another -- but my back-of-the-envelope optimal
partisan gerrymandering strategy that includes this dynamic is still to set
all opposition districts to the same level of partisan strength unless the
number of opposition districts equals 0 or 1. There may be other geographic
factors that intrude, for example, the two Republican leaning congressional
districts in Maryland are located in the DelMarVa peninsula and the
panhandle. There is no easy way to shake up these districts. There are also
limits as to what can be done to voting rights districts.

============
Dr. Michael P. McDonald
Associate Professor, George Mason University
Non-Resident Senior Fellow, Brookings Institution

                             Mailing address:
(o) 703-993-4191             George Mason University
(f) 703-993-1399             Dept. of Public and International Affairs
mmcdon@gmu.edu               4400 University Drive - 3F4
http://elections.gmu.edu     Fairfax, VA 22030-4444

-----Original Message-----
From: Antoine Yoshinaka [mailto:antoine.yoshinaka@ucr.edu]
Sent: Saturday, October 30, 2010 1:52 PM
To: Michael McDonald; 'Election Law'
Subject: Re: [EL] Redistricting answer

One reason why there is an upward limit to competition may be due to
partisan segregation.  Bruce Oppenheimer has some work showing that more and
more Republicans live near Republicans and Democrats near Democrats.  So it
may be difficult to draw districts with a good number of both R and D
voters.

As for the finding that optimal partisan gerrymandering leads to safe
districts, I would like to add one wrinkle.  Chad Murphy and I published a
couple of articles (2009 Political Geography and forthcoming Political
Research Quarterly) in which we show that partisan mapmakers are able to
target out-party districts in a different way, by inducing population
instability.  So for instance, it may be the case that an out-party
Republican incumbent's district is composed mainly of Republican
constituents both prior to and following redistricting (i.e., it's a safe
district in terms of partisanship).  However, when many "old" Republican
constituents are replaced with "new" Republican constituents, it creates an
environment propitious to a competitive election.  That's what we found
happened in 2002.  By looking mainly at partisanship, optimal gerrymandering
models potentially miss an important source of heterogeneity which leads to
competition.

Antoine
--------------------
Antoine Yoshinaka
Assistant Professor
Dept. of Political Science
2217 Watkins Hall
University of California, Riverside
Riverside, CA 92521
Tel:(951)827-4688
Fax:(951)827-3933
Email: antoine.yoshinaka@ucr.edu
Website: http://politicalscience.ucr.edu/people/faculty/yoshinaka/index.html

---- Original message ----