Not really sure why this article from May is suddenly relevant on hn, but anyway:
While gerrymandering does help explain the GOP's outsized representation in the House relative to their national vote share, the increasing dominance of the Democratic vote in large cities is probably more significant[1]. Essentially, if you wanted GOP representation in the House to mirror their national vote share, you'd be forced to draw some pretty nonsensical districts that sliced up urban centers so they could be mixed with the outlying (and sometimes way outlying) suburbs. Getting 90%+ of the vote in ward after ward of Chicago doesn't help your guy in Deerfield when he loses by 2%.
This is exactly why the GOP probably won't lose the house anytime soon, but national elections almost always go to the democrats. The GOP will probably pick up a bunch of senate seats this year because of low turnout and an abnormally strong anti-incumbent bias, but they have almost no chance at winning the presidential election in a few years.
At this point, the parties have devolved to mirror the culture wars that have beset this country since its founding: urban vs. rural. There's money on both sides - agriculture and mineral wealth in rural areas, and finance/industrial wealth in urban areas, so it's not a rich vs. poor thing. It's almost purely a sociological divide along relative geographic lines at this point.
Then the GOP has done an exceptionally good job of selling that they're about less government control (when really, their stance on government intervention depends on the issue).
Just look at all the GOP talking heads saying we need more restrictions on travel from Africa, we need to restrict the definition of marriage, we need to prevent abortion, we can't cut military spending, we need a huge border patrol, etc. The GOP beats the drum of small government, but are the first to enact big government when it suits their needs. See the Department of Homeland Security, enacted by a republican president with a republican congress, and which has in 12 years grown to become one of the biggest and most wasteful government agencies out there.
Politicians lie. It's their job. So we can't judge them based on what they say, we have to judge them based on what they do. And right now, unfortunately, an individual politician doesn't have the ability to contradict his party if he wants to keep his job. So we have to look at the party's track record, and when we do, it's pretty obvious that there's no choice for "relatively less government control". Both parties want more government control; just that the democrats want it in the form of government bureaucracy that creates dependency on the system, and the republicans want to dictate what people can and cannot do, then enforce it with an increasingly militarized police force.
So, you're against "government control," but you support government control of abortion, state-sanctioned persecution queer people, and government control of movement across borders?
Let me guess: You're a white heterosexual cisgendered man.
Why is 'compactness' valued? If a district should be people of a similar background/interests, then the geography is messier- the NC 12th district is basically the I-85 corridor, which may make more sense than a square of arbitrary size. If a district is just supposed to be a random container of voters to apportion representation, then base it on a voters' SSN or something and remove geography entirely.
Well, the state boundaries themselves are gerrymandered in some cases. Do you think it's an accident that a straight line through Kansas City divides Missouri and Kansas?
At some level, all boundaries are arbitrary. Not to say they couldn't be more just and logical, but it's not like the Texas panhandle is some sort of platonic ideal.
I've long wondered how much you could reduce (obviously not eliminate) gerrymandering with a minimum compactness requirement for the shape of each district. Or maybe a requirement that all districts must be convex polygons (with allowances for concave state borders)? I think convex polygons map pretty closely to what most people expect districts to look like, and it prevents compact but still problematic "fat horseshoe" shaped districts.
I have a hard time figuring out if this is just a case of me reaching for my mathematical hammer, or if this would actually help.
It's probably an overcorrection to never take into account whether the people represented by a seat have something in common. A cartoon example would be 4 counties arranged in a square, where one of the counties has a city that is roughly equal in population to the 3 other counties combined (or so). 2 equal shaped districts, with the city chopped in half, is not obviously a better arrangement than a district mostly comprised of the city and one mostly comprised of the outlying areas.
I agree that "voting history" isn't a great thing to be considering as something people have in common (at a minimum, it shouldn't be a high priority).
I was thinking more that compactness and/or convex polygon-ness could just be necessary constraints, not sufficient. You'd still get to draw the lines in ways that split the state's population according to shifting population and demographics, but you wouldn't be able to draw districts that are quite so crazy looking.
It's worth pointing out that Maryland as a whole would probably score poorly on the metric they're using because of geography. There's no excuse for the 3rd District, but because it's a small place with a lot of water, the other districts look worse than they are because of having to work around the coastline.
I agree. The analysis seems weak. Population growth isn't uniform across the state, so why would districts be uniform? You'd constantly need to tweak each one to keep the same number in each. Unless you want to completely redraw all the districts each election.
"In those six states, Republicans picked up about 11 more seats than you'd expect from simply looking at the parties' vote shares."
The fact is, we don't decide representation based on state-wide vote share, so it's not a good proxy for "how it should be".
Keep in mind, I'm not supporting the current districting process. It should be non-partisan and the folks getting elected should have no say in how they are drawn.
If you squint, some of those districts look like Julia sets. But I guess that if you were selecting districts that have low compactness you would necessarily be selecting for districts that look fractal-like.
I never understood why this is an acceptable thing to do. It's democracy except you get to pick who votes for you and who doesn't, so, it's actually not.
It's a democratic republic, to be specific. Republicanism is the form of the government (representation of the people), democratic elections are the mechanism of selecting those representatives. You could have, e.g. a technocratic, plutocratic, or a constitutional monarchial republic.
While gerrymandering does help explain the GOP's outsized representation in the House relative to their national vote share, the increasing dominance of the Democratic vote in large cities is probably more significant[1]. Essentially, if you wanted GOP representation in the House to mirror their national vote share, you'd be forced to draw some pretty nonsensical districts that sliced up urban centers so they could be mixed with the outlying (and sometimes way outlying) suburbs. Getting 90%+ of the vote in ward after ward of Chicago doesn't help your guy in Deerfield when he loses by 2%.
[1] http://mobile.nytimes.com/2014/09/07/upshot/why-democrats-ca...