538's take on gerrymandering (US politics)

Since no one is going to gerrymander for partisan competitveness, the only fair system on the table is compact districts. Which gives the GOP a +28 to +30 advantage, compared to their current +34 advantage.

Nate Silver’s 538 did several alternative Congressional maps at the above link, ranking them by competetiveness, compactness, following existing borders, and a few other characteristics. There was no map that scored very highly on all of these characteristics.

Which map is best, and why?

How could this be instituted?

For the first question, I think I like the map made for compactness that tries to follow existing borders slightly better than the map that maximizes competetiveness – the latter results in several ridiculously shaped districts, and is therefore a sort of gerrymander of its own, even if it’s a non-partisan gerrymander.

For the second, I assume the only plausible way would be a Constitutional amendment. That would be extraordinarily difficult, since most in Congress would be loathe to make their own re-elections less likely.

First, this and adahar’s thread on the same 538 article might be best merged.

Into the the meat … there are some good historic reasons to apply some “majority minority” principles and the “compact” versions while having the advantage of being somewhat arbitrary are noted as possibly violating the Voting Rights Act.

As adahar’s take highlights though, the exercise showcases that the GOP House structural advantage is not exclusively due to partisan gerrymandering … which has been pointed out here before. That’s best seen by the 538 “efficiency gap” column which shows the size of the partisan advantage of each map and which party it favors. Unless one moves into the explicitly political result motivated gerrymanders of “competitive” or “Democratic gerrymander” there is a structural GOP advantage (or Democratic handicap if you prefer) for the House. That is something the Democratic side will just have to deal with.

I think the Voting Rights Act mandates some consideration of majority minority principles and “best” is having that modify a basic approach based on compact by borders, created by non-partisan committees. The easier lift though is eliminating the aspects that are most egregiously gerrymandered with partisan intent.

Coming from a Country with no gerrymandering, because that would be against the constitution, I think they start with the wrong metrics.

Why maximize for “competitveness”? In what way is that useful to meet the criterions for truely democratic elections (equal, free, secret, re-checkable, not-falsifiable)?

Since I think their starting conditions are wrong, None.

If the US ever gets over their “we are the best in the world, and unique in the world”, they could just ask the rest of modern democratic countries how to draw districts. You guys don’t have to re-invent the wheel, just copy an existing one.

But since gerrymandering, together with voter Suppression and majority-only voting System favours the big Party, especially the one in power, I see any changes at all unlikely.

E.g. an overhaul of the voting System for me would also contain the right to a national - not state! - photo ID, issued without a fee at not just the DMV, but also post Offices and many other official institutions, conforming to modern security Standards. To apply, not a copy of birth certificate is necessary (since this is difficult for People in US), just a state Picture ID /license maybe with Social Security number.

This ID, with one national design, performing to current safety Standards, would also make all the other Tasks easier where 50 different designed (and easily forged in China) driver’s licenses are used (e.g. buying alcohol).

Similar thread http://boards.straightdope.com/sdmb/showthread.php?t=847810

I’ll defer to that one then, thank you.

If 538’s numbers are correct, then there really isn’t much of a gerrymandering problem. The GOP has a six seat advantage over compact districting. I think the courts are going nuts in part because they aren’t measuring GOP gerrymandering against that metric, but against the metric of how things were before, when Democrats gerrymandered to their advantage. That actually has caused a pretty wild swing compared to how it was from 2001-2010, and no doubt a HUGE swing over the period of Democratic House dominance that prevailed up till 1994.

nm

Replying to your post in the other thread…

But that +30 advantage is the difference between 151 D and 180 R, with 104 competitive districts. Compare that to the current system that is 168 D and 195 R, with a third fewer competitive districts.

Why is this important? Under the first scenario, with 180 solid R seats, they must take 38 out of 104 (or 37%) competitive seats to gain a majority. Under the second scenario, they must take 23 out of 72 (or 31%) competitive districts to gain a majority.

In other words, would you rather have to win 31% of open seats to gain a majority? Or 37%? Either way, the map is tilted toward Republicans, no doubt, but the paucity of competitive seats does make a big difference.

The problem is, neither party is going to ever be interested in having 180 competitive seats. If there’s one thing they both agree on, they want mostly safe seats.

Also, 538’s numbers average everything out. Even with compact districts, the GOP would still be expected to win +28 seats as a median.

I think an argument can also be made that 180 competitive seats might be LESS representative of voters in some ways. It would give a rather large balance of power advantage to Blue Dogs and liberal Republicans. Which in many ways would be a good outcome, but would leave both party bases in a very dissatisfied state.

Yes, there’s no way to get around the fact that less populous states that lean red, like Wyoming and the Dakotas, give a permanent advantage to Republicans of a handful of seats, and that advantage cannot be erased through the drawing of districts in other states.

But what are your thoughts on drawing districts on the proportionally partisan basis? That would seem to have the advantage of both increasing the number of competitive seats, eliminating Republican efforts to minimize minority representation in Congress, and minimize the so-called efficiency gap. Seems like a reasonable starting point.

I do not think they are measuring against the (relatively more weakly) Democratically gerrymandered maps. Of course we will see what the Supremes do accept, or refuse to accept, as a metric to judge against, but “proportionally partisan representation” is possibly the fairest baseline to judge against.

If you’re going to go that route, constanze, then the goal is to maximize representation. So then the point would be to come up with a system that matches as closely as possible the percentages of each party in the populace with the number of people in each party in representation.

The only system I’m aware of that does this is not put in place anywhere in the world (as far as I know). But it is called Proportional represenation. Here is a nice video that explains it simply.

As you say, it isn’t possible that we would fix this. But you do miss that there can be a discrepancy between the national and state level, which could make some sort of compromise possible. Right now, removing gerrymandering helps, but the real solution is to gerrymander in such a way as to get proportional representation. It’s not about competitiveness, but about making sure people have representatives in Congress.

I agree it’s fair, but I don’t like the idea. Besides, that’s already built into the system. They already draw districts so as to benefit a party, they just do it by taking turns rather than through cooperation. I’d actually rather see geographically compact district drawn by computer algorithm.

But if they admit 538’s data into the testimony, they find that at least nationally there isn’t really much unfairness going on. The democratic system seems to be working well enough. At least well enough that the courts don’t really need to get involved.

You’re assuming a national analysis, adaher. And that doesn’t fly. Each state is considered its own beast, here. The courts have to look at how representation is maintained on a state basis. So just because nationally it balances, that doesn’t make Pennsylvania or North Carolina any more constitutional.

Similar threads merged.

Well if you agree (as you just did) that proportional representation would be the most fair (even if you don’t like it) then there is a lot of unfairness going on. It is just not primarily a direct result of partisan efforts to be unfair.

ISTM that extremely convoluted exercises that attempt to create significant partisan advantages should be disallowed even if the impact of such efforts end up being only of moderate impact. We do not make shooting a gun at someone with intent to kill not a crime just because the shooter usually misses.

I actually prefer geographically compact districting, with some modification for keeping like-with-like (for example, a district that is mostly urban and a district that is suburban to rural is preferable to two districts that are all three). But the trouble with compact districting is that it reduces the “value” of a vote in areas where there is more unanimity of voting, such as urban settings in most of America, which tend to be very Democratic. The fact that Republicans are currently spread out all over the place (relatively speaking) gives them an advantage in any system that compacts districts geographically.

I’m fine with that (it would encourage the Democrats to adjust their politics to stop being the party of the urban masses), but politically, that’s never going to fly. And, of course, there is the question of how to police “compactness” so that it can be measured in a way that lets courts decide when districts aren’t being “compact” enough.

True. If PA was indeed gerrymandered to give Republicans an unnatural 3 to 4 seat advantage, that shouldn’t stand.

But according to 538, the GOP gerrymander is actually less than one seat:

Currently the GOP enjoys a 6 seat advantage, when a compact district scheme would generate than a roughly 5 seat advantage. PA is really indicative of how the nation’s geography is stacked against Democrats. This is a pretty blue state, but even under a “fair” system, the GOP should have twice as many seats as the Democrats.