What with the recent spectacle of Tom DeLay waltzing into Texas (in an off year) and breezily eliminating six or seven Democratic seats, coupled with the new enthusiasm for the Schwarzenegger-backed “fair redistricting“– I’ve been thinking about gerrymandering. A modest proposal:
For any proposed congressional district, take the perimeter squared and divide by the area. This number yields the district’s gerrymander number, G.[1] Now set a cap on G — say, for the sake of argument, 100.
Taa-daa.
Oh, and another thing — if I were King, “team” would be spelled T-E-I-M. Thank you very much.
1. For example, a circular district would have a G = 4pi, a square district would have a G = 16, a 4 x 1 rectangular district would have a G = 25, and so on.
Does this assume that districts made of simpler shapes would be fairer? Wouldn’t this potentially adversely affect coastal regions and other places where the lines CAN’T be drawn straight? (And don’t say that you would drawn the lines over the ocean, that’s federal territory. :P)
I think the shape of a district matters less than where geographical the lines on the map fall. There’s no rule that a district shape has to be funky in order to draw lines so that Party XYZ always has a majority. You just makes bunch of squares and place them so the lines always dissect the opposition party’s stronghold communities.
My off the cuff starting point is Area codes. First divide the state into voting “regions” by area codes. Then draw the minimal number of straight lines (using major streets or other physical markers whenever possible) across the area codes in order to equally subdivide the regions into districts. I agree that if you do this repeatedly, the phone companies will catch on and exploit the arbitrary descretion that is afforded them, but as an initial measure for more or less impartial districts, I can’t think of a better starting point.
(The reason why I picked area codes is that they are based loosely on population sizes so they could lend easily to being used.)
Molly Ivins, speaking on NPR, brought up Iowa as an exercise in redistricting sanity. Those killjoys decided to take politics out of redistricting by appointing a non-partison panel (gasp!) to determine the districts. This worked exceedingly well and has led to close races, which means that those running for office have to pay attention to the voters in their district.
Speaking of whom, I went to an Ivin’s book signing event last night with a friend and purchased her new book *Bushwhacked.* We got there an hour ahead of time. All seats were taken and people were already standing in the aisles. There must have been 250 people there by the time she began speaking. This would give me hope, but, then, the Bay Area didn’t vote for Arnie either and look how far that got us.
Mom
I like the idea, although it should be noted that many of the most obviously gerrymandered new districts in Texas (especially the “stripe” districts) would have relatively low G values compared to some classical gerrymandered districts with tendrils reaching all over the place.
Not a bad proposal, though the issue of how to do districts where straight-line borders aren’t possible is an issue.
I would propose resolving that thusly. When a given edge for a district’s border EXACTLY follows either a natural boundary (rivers, mountain ranges, interstate highways but no other roads, etc), a county line, or the state’s border, then for purposes of determining the gerrymandering number the length of that edge will be treated as though it were a straight line between the endpoints of that edge.
The idea behind this is to keep the idea of rewarding straight lines and other simple shapes (which are almost always the most rational way to determine borders) without penalizing those situations where a straight line is not possible (because of county or state borders), or where some major natural feature would make more sense than a straight line.
Adam: You may be right, this proposal only works up to a point for “stripe” districts, depending on where you set the threshold. It really only works well on the traditional “octopus” gerrymandered districts. But it’s a start!
Millenium (and Adiv): You’re right, it’s unfair to penalize districts that fall along natural boundaries. I think your idea is sound, as long as one defines very carefully what a natural boundary is. That’s what lawyers are for.
I’ll get right on it, chief!
[Since I am Evan’s Brother-in-Law (school), I take the last comment to be at least partially directed at me.]
I have another idea, which I posted to my blog awhile back (I’m a Texas resident) but reposted by chance tonight due to a blog software issue… basically, my answer is to let the people gerrymander themselves.
I have another idea, which I posted to my blog awhile back (I’m a Texas resident) but reposted by chance tonight due to a blog software issue… basically, my answer is to let the people gerrymander themselves.
Sorry for the double-post: the permalink for that post is linked to my name.
I’m shocked at such a display of poor spelling: the proper spelling of team would be T-I-E-M. “IE” in a work is pronounced “eeee” while “EI” is pronounce “aye”. Tish.
Wow, that’s fascinating! You see, in my native land, the obscure, exotic state of California:
– We pronounce “IE” with an “aye” sound (as “pie”, “tied”, and “lie”).
– We pronounce “EI” with an “eee” sound (as in “belief”, “Leif Erikson”, “conceit”, and “being”)… unless the “EI” is followed with a “GH”, in which case it takes on an irregular “long A” pronunciation.
How upsetting to hear that we’ve had it backwards all these years.
A better way to reduce the effects of gerrymandering would be to have larger districts, and elect several representatives at a time from each district, using one of several available “proportional representation” methods. Implimented correctly, such multimember districts are much harder to gerrymander. And proportional representation would bring us many other benefits.
See
http://fixour.us/
or
http://fairvote.org/
PR would have a bunch of benefits, indeed, but I’d favor using it for only one of the two legislative houses, probably the upper one. (I’d also favor reforming the Federal Senate to have about then Senators for CA and only one for WY — that way Wyomans would only have seven times more Senate representation, per capita, than Californians, instead of their current seventy.)
The problem with a straight P^2/A ratio cap is that you have places, such as areas of coast that are cut off from inland population centers by mountains, where a long thin district makes good sense — a community of common interests lives along a long snaky area between the mountains and ocean. You can use the mathematical rule in Kansas or Texas, but in California, it’s better to use logical rules: A district may have no more than six edges (or five, or eight, or whatever — some samll numer), and an edge is defined as a coastline, river, political boundary (city or higher), or numbered non-residential route.
Heck, forget proportional representation — is there anything in the U.S. Constitution that says California can’t suddenly switch to winner-takes-all? “I’ll see your illegal off-cycle gerrymandering, and raise you…”
As for the P^2/A cap — yes, I suppose there could be some pathological cases. Your “edges” plan is good, although I think it also allows a bit more wiggle room for gerrymandering, I think. The definition of those edges sounds simple, but I bet someone would figure out some way of suing over the *exact* definition.