Gerrymandering?

[marcusmarcusrc]

Those of you who are my facebook friends may have noticed some posts by me about various voting issues - bemoaning my lack of Senator or Representative, for example - but also pointing out what I think is an interesting paper by a family friend of mine on gerrymandering (google stephanopolous gerrymander to see a few of his articles).

The recent election, and the Republican majority in the house, demonstrates the results of gerrymandering quite nicely: http://thinkprogress.org/justice/2012/11/07/1159631/americans-voted-for-a-democratic-house-gerrymandering-the-supreme-court-gave-them-speaker-boehner/?mobile=nc. I did a quick calculation on Pennsylvania, and determined that though a majority of Pennsylvanians voted for Democratic House of Representative candidates (50.7% to 49.3%), the state as a whole elected 13 Representatives to 5 Democrats.

One question I have is that, while the PA gerrymandered districts look ridiculous, would fixing them actually solve the disproportionate representation problem? After all, urban centers vote 80 to 90% Democrat, and most mapping schemes, whether by geographic compactness or by "spatial homogeneity" (see the Stephanopolous Law Review article), are going to keep the urban area as an intact entity. So I think even a neutrally designed PA map will end up having a Republican advantage. Another option that Stephanopolous raises is multi-member districts - this seems like it would help improve the fidelity of representation (and could possibly be combined with some kind of preferential voting system), make third parties more viable, and reduce extremism.

Thoughts?

Comments

[dcltdw]

Wouldn't you have lots of small districts in cities (because population density is high) and very large districts in the rural areas? Think concentric circles -- the opposite being the PA map (spokes on a hub centered on cities).

[marcusmarcusrc.livejournal.com]

Yeah, you would definitely have small city districts and large rural districts, but the small city would have, say, 3 districts each with 90,000 D votes and 10,000 R votes each, and the large rural area might similarly have 3 districts with, say, 60,000 R votes and 40,000 D votes each... and so you split the state 50/50 R/D even though the population is almost 2:1 D:R...

[marcusmarcusrc.livejournal.com]

Oh... or are you pointing out that the PA map is not small-urban/large rural, because they are merging urban/rural areas in the gerrymandering process? There might be some of that. That's certainly what they've done in Utah, where the Salt Lake Urban area is split between 3 large districts each covering about a quarter of the state. Fortunately, the D managed to pull off an upset victory anyway.

So, yes, compactness might be an improvement over the current gerrymandering, but I'm not convinced that a purely rational, compact map would actually lead to a fair final outcome. Though part of the problem here is the two-party system - if it was a jungle-election, you'd get a large number of moderate Republicans in the 60/40 states, and some very liberal Democrats in the 90/10 states, and summed together that might appropriately represent the state. But instead, what you seem to often get is a system where in order to win the primaries, you can't have real moderates, and by the time they get to the general election, the only thing that matters is the letter next to their name so a moderate D can only occasionally pick off an extreme R even in a 60/40 district...